The October Connection

[tacoman25]

I originally posted this in the New England winter thread, but figured it made sense to post it here as well, since it applies to just about everyone.

I think this illustrates the October connection during -ENSO years pretty well. The first map is all -ENSO Octobers that were clearly colder in the East than the West overall. The next map is the following winters. Then we have all -ENSO Octobers where the West was clearly colder than the East. And the following winters.

[BethesdaWX]

Seems like a good connection signal to watch. Can you do a map seperating the PDO phases?

[tacoman25]

Seems like a good connection signal to watch. Can you do a map seperating the PDO phases?

By phases...

-PDO/warm western October

-PDO/warm eastern October

+PDO/warm western October

+PDO/warm eastern October

[Lee]

Interesting comparisons Tacoman. Although, there were still some years that went "against the odds", so to speak. '95-96 sticks out like a sore thumb. 70-71 quite a bit as well. 75-76 did Dec. and Jan. and then blow torched Feb. Even the mid-80's were quite cold in much of the east. Take away a couple of "super warm" years,;I.e, '49-50, 73-74 and 07-08 and the evidence is not so great. Likewise for the colder east comparison maps.

So, it's still not that great of a deciding factor,i.m.o...but still somewhat interesting results.

Btw, not meaning to be picky,lol, but you left 50-51 out again. It was a -enso year.

[tacoman25]

Interesting comparisons Tacoman. Although, there were still some years that went "against the odds", so to speak. '95-96 sticks out like a sore thumb. 70-71 quite a bit as well. 75-76 did Dec. and Jan. and then blow torched Feb. Even the mid-80's were quite cold in much of the east. Take away a couple of "super warm" years,;I.e, '49-50, 73-74 and 07-08 and the evidence is not so great. Likewise for the colder east comparison maps.

So, it's still not that great of a deciding factor,i.m.o...but still somewhat interesting results.

Btw, not meaning to be picky,lol, but you left 50-51 out again. It was a -enso year.

Yeah, with any correlation not every year is going to follow it. It's just about odds.

I didn't included 1950-51 and a number of other -ENSO years because there was no clear October signal.

[tacoman25]

To further examine the correlation, let's look at what percentage of years actually follow the October signal.

Here are the years with the clearest warmer west/cooler east temperature pattern: 1903, 1906, 1909, 1917, 1933, 1944, 1964, 1967, 1974, 1988, 1999, 2008, 2010.

1903, 1906, 1917, 1933, 1944, 1964, 1967, 2008, and 2010 all were followed by winter temperature patterns that were cooler in the East compared to the West. The correlation worked for 69% of years.

On the other side, here are the years with the clearest cooler west/warmer east temperature gradient for October: 1908, 1916, 1949, 1954, 1956, 1970, 1971, 1973, 1975, 1984, 1995, 1998, and 2007. The correlation in winter followed for 1908, 1916, 1949, 1954, 1956, 1971, 1973, 1984, 1998, and 2007 (1970-71 was borderline), or 77% of the years.

This makes the overall correlation/inverse correlation percentage average: 73%. With a sample size of 26 years, that's pretty good.

If anyone has any questions about any of these years I can provide the maps demonstrating the temperature patterns referenced.

[HM]

Let's give this here correlation some theory!

So after further investigation into the cold ENSO-October-winter connection, I think that not only is there some validity to the correlation but it also seems to fit into my qbo/forcing thoughts in the pinned thread.

Since the idea is based off trough/ridge placement, I had to reanalyze the years' 500mb charts and strictly from 1949-50 on so I can use modern variables. I went through the list presented in this thread and others and picked years that were solidly in one camp vs. another (not too many changes from taco's original list). Here is that information:

Cold ENSO Octobers with a solid PNA pattern (10):

1955

1962

1964

1966

1967

1974* borderline rejected

1988

1999* borderline rejected

2008

2010

Their 500mb composite:

Cold ENSO Octobers with a solid RNA pattern (11):

1949

1950

1956

1970

1971

1973

1975

1984

1985

1998

2007

Their 500mb composite:

Rejected years because the monthly average wasn't clearly in one camp:

1954

1959

1960

1961

1989

1995

1996

2000

You run the two camps forward into their DJF mean and they look like taco's finding (and my QBO composites too):

PNA-October's DJF 500mb composite:

RNA-October's DJF 500mb composite:

So after looking over everything, it hit me this morning how everything between taco's correlation and mine could work together. Yesterday, I expressed to taco that in October especially, tropical forcing/MJO activity plays a key role in the overall PNA or RNA state.

To see MJO temperature composites, go here: http://www.cpc.ncep....es/Temperature/

Simply, forcing that focuses near the "El Niño phases" or phase 7-8-1, which is near the dateline, roughly, allows October waves to position for a general RNA pattern. It can be counterintuitive because everyone is used to these phases bringing more of a PNA pattern during the winter. But in the fall, waves are still too short for that to work. Forcing that occurs from the Indian Ocean into the Tropical West Pacific (phase 2-6) will generally focus on a cool eastern US picture during the autumn. Years that were rejected likely had out-of-place forcing against the stagnant ENSO forcing or a very pronounce propagating MJO wave that created significant week-to0week pattern variations.

But trying to figure out all of these forcing processes/circulations and MJO activity is not only time consuming but likely a waste when trying to explain the winter connection. So, I tried a different approach:

If you simply plot OLR anomalies, you will only have years strictly from the 70s on; and on top of that, they will all show the same general OLR pattern that you would expect with a cold ENSO event. So this will fail, especially on a monthly average of 10 or so years. After going through a lot of the typical parameters and coming up short, I finally plotted their 50mb winds. Initially, I didn't want to do this because I knew that the years looked mixed to me at first glance. But, to my surprise, here's what they show:

PNA-October 50mb zonal winds:

RNA-October 50mb zonal winds:

Well what do you know, there seems to be more westerly 50mb winds in the PNA October years and more easterly 50mb winds in the RNA October years. The theory goes: +50mb winds in tandem with cool ENSO will reduce convection further in the El Niño phases and help increase the off-equator convection. This sets up the forcing composites that favor cold in the East during an October. Meanwhile, -50mb winds would do the opposite. To see this on a map is tricky. Like I stated earlier, if you plot OLR you are going to run into trouble. So, I decided to look at UVM for convection and infer circulation.

PNA-October 500mb UVM:

RNA-October 500mb UVM:

While not significant, probably due to the mixed qbo/ENSO strength years within the composites, the -50mb/RNA years have more uplift near the Dateline and more subsidence across the tropical East Pacific than the +50mb/PNA years. This fits with the forcing ideas above, perhaps allowing the RNA October years to have more of a tendency to get the forcing into the El Niño phases. Also, pay attention to the North Pacific and how z-circulations are being influenced by these subtle changes, too.

It is my belief that the October temperature pattern may indirectly be trying to tell you how the stratospheric-tropospheric connection is behaving. Most likely if your +QBO or -QBO winds are solid enough in one camp, they will likely carry into the next two months or so affecting convection. Even if they are waning, they are probably still affecting it more than the oncoming phase. If you take these processes and carry them through the late autumn/early winter, a few things will happen:

1. You will setup a greater Aleutian High in the +50mb years.

2. You will have a tendency for cyclonic wave breaking in the West Atlantic capable of taking the North Atlantic High (see October PNA years) and displacing it into Greenland.

Between the Aleutian High and North Atlantic High, enough waves will come down into North America that will cause a blocking pattern at some point to develop. All of this is heightened if other factors also favor blocking, too.

So for the helluva it, I ran back my personal QBO years to see what they looked like in October. Well now, isn't that something. They look like taco's October maps!

HM's westerly QBO / cold ENSO winters' proceeding Octobers

HM's easterly QBO / cold ENSO winters' proceeding Octobers

The QBO system in tandem with ENSO doesn't account for it all, that's for sure. But it certainly seems to bridge a few of the concepts being thrown around, along with some correlations.

[VAwxman]

Good stuff. Looks like that, unfortunately, this October's upper air pattern isn't going to fit well into either camp.

[HM]

Good stuff. Looks like that, unfortunately, this October's upper air pattern isn't going to fit well into either camp.

True! Glad this stuff is being discussed and ironed out during an October that may not give us a clear signal. The only thing amazing is that the September 50mb QBO actually went up to +10, which is now the new peak. Pretty impressive!

[tacoman25]

Thanks, HM. A lot of this makes sense, though the interesting thing is that even though your QBO year composite matches my composite, some of the years are totally different - for example 2008 was a PNA match, but the QBO was quite different. But of course, not every year is going to fit with theory.

[tacoman25]

Good stuff. Looks like that, unfortunately, this October's upper air pattern isn't going to fit well into either camp.

With the way the LR is trending, I think it's more likely going to end up looking like the PNA composite. Which would be good news for those in the east.

[HM]

Thanks, HM. A lot of this makes sense, though the interesting thing is that even though your QBO year composite matches my composite, some of the years are totally different - for example 2008 was a PNA match, but the QBO was quite different. But of course, not every year is going to fit with theory.

Take a look at 50mb though:

2008 -14.94 -12.65 -9.84 -6.67 -1.79 4.45 7.49 7.23 6.65 7.30 6.91 7.73

2009 8.35 8.19 7.07 9.74 11.34 9.13 4.00 2.84 3.69 4.49 3.51 1.44

2008 was a classic +50mb autumn and winter.

[VAwxman]

With the way the LR is trending, I think it's more likely going to end up looking like the PNA composite. Which would be good news for those in the east.

I don't really have a good handle on it. It could really go either way and will likely be determined by the last week to ten days. FWIW my guess is we wind up quite mixed with warm in the middle rather than east or west.

[tacoman25]

To illustrate the possible winter implications, here again are the DJFs following HM's +PNA Octobers and -RNA Octobers.

[tacoman25]

Take a look at 50mb though:

2008 -14.94 -12.65 -9.84 -6.67 -1.79 4.45 7.49 7.23 6.65 7.30 6.91 7.73

2009 8.35 8.19 7.07 9.74 11.34 9.13 4.00 2.84 3.69 4.49 3.51 1.44

2008 was a classic +50mb autumn and winter.

That's true. I know you discussed this some in your other thread, but which level of the QBO do you think is most important?

[HM]

That's true. I know you discussed this some in your other thread, but which level of the QBO do you think is most important?

In the case strictly of what I was discussing in the other thread with how the circulation cells work in the North Pacific...50mb/lower levels hands down. Every case that was either classified as westerly or westerly in the lower levels (50mb) had a poleward Aleutian High (DJFM average). Everyone of them.

[tacoman25]

I don't really have a good handle on it. It could really go either way and will likely be determined by the last week to ten days. FWIW my guess is we wind up quite mixed with warm in the middle rather than east or west.

Very possible. Well, if that is the case our best analogs might be....

Which led to...

[VAwxman]

Very possible. Well, if that is the case our best analogs might be....

Which led to...

Now THAT October map looks closer to what I'd expect for this month as a whole.

[HM]

Taco, didn't 1973 and 1956 also have the same kind of look? I may be missing some others but I thought I remember those two years looking like that central us warm composite.

Edit: You could make an argument that 1996 and 1989 also fit from the rejected years listing. By adding in either 73/56 or all 4 years, you increase the warmth in the Southeast but the Upper Midwest down into Southwest look cold. Typical gradient look over the Northeast.

[phlwx]

this is a good thread...good stuff tacoman, hm, etc...