QBO relationship to the Arctic Oscillation

[OKpowdah]

Has been a pretty clearly understood relationship, but here's what I've found with my re-definition of the QBO in phase space:

 

Phase -90 corresponds to an easterly QBO and phase +90 corresponds to a westerly QBO.

 

 

Notice that the strongest association to a -AO occurs not at the QBO min, but actually closer to the easterly shear phase. Same with a +AO, the strongest association is closer to the westerly shear phase

 

[OKpowdah]

Here is the 500mb height correlation for phases 30 and 210 (-150)

 

 

[Mike.Ventrice]

Hi Guys,

 

Created a phase space for the AO index (similar to CPC's AO but I created my own EOFs)

 

 

Forecast is for phases 7->1. Composites show this typically has warm northern plains, cold Southeast/Mid Atlantic... Not a bad match for what should occur this upcoming week/two.

 

[HM]

Has been a pretty clearly understood relationship, but here's what I've found with my re-definition of the QBO in phase space:

 

Phase -90 corresponds to an easterly QBO and phase +90 corresponds to a westerly QBO.

 

Screen Shot 2013-04-23 at 2.34.11 PM.png

 

Notice that the strongest association to a -AO occurs not at the QBO min, but actually closer to the easterly shear phase. Same with a +AO, the strongest association is closer to the westerly shear phase

 

Thanks man; I was waiting for you to tackle the AO with your new index. I have a few questions: why did you choose to use DJFM? Also, what about prior research that seemed to indicate that the QBO's influence is heaviest at the start of the winter, particularly during the radiative phase of the stratospheric vortex? Would your numbers still be significant if you only used the first two months? What if you did OND or NDJ?

 

Perhaps none of this is necessary if you feel certain shear stresses / MQI phases dictate the overall season's AO character and that's all you are trying to show.

[OKpowdah]

Thanks man; I was waiting for you to tackle the AO with your new index. I have a few questions: why did you choose to use DJFM? Also, what about prior research that seemed to indicate that the QBO's influence is heaviest at the start of the winter, particularly during the radiative phase of the stratospheric vortex? Would your numbers still be significant if you only used the first two months? What if you did OND or NDJ?

 

Perhaps none of this is necessary if you feel certain shear stresses / MQI phases dictate the overall season's AO character and that's all you are trying to show.

 

Honestly, I was working on this on the side of a lot of other stuff and didn't even think much further than I just wanted to check it out for the 4-month winter season. So you make an excellent point here. 

 

I just ran the same stats for every trimonthly period. I wouldn't want to try any less than three month periods because of sample size issues for each phase.

 

So for each of these periods (centered on the month listed), I calculated the 33% and 67% percentile thresholds (variable). I also used the full year percentile thresholds for comparison (constant).

 

 

 

So the strongest influence (with clear 180-degree association) appears in the NDJ period and still pretty good for the DJF period. The 180-degree association then starts to break down (i.e. maxima and minima not occurring in opposite phases).

 

But FMA is really interesting to me, because there is a MAXIMUM, and a big max at that, in the EASTERLY phase of the QBO ... for phase -60, ZERO months below 33 percentile. This is followed by a rapid switch to a big minimum in late spring (AMJ) ... for phase -30, ZERO months above 67 percentile. That's pretty significant to me. And I'm really curious about that switch during the spring season.

[DaculaWeather]

That's great Mike! 

[am19psu]

Awesome work, guys. Looking forward to more QBO secrets.

[HM]

 

Honestly, I was working on this on the side of a lot of other stuff and didn't even think much further than I just wanted to check it out for the 4-month winter season. So you make an excellent point here. 

 

I just ran the same stats for every trimonthly period. I wouldn't want to try any less than three month periods because of sample size issues for each phase.

 

So for each of these periods (centered on the month listed), I calculated the 33% and 67% percentile thresholds (variable). I also used the full year percentile thresholds for comparison (constant).

  

So the strongest influence (with clear 180-degree association) appears in the NDJ period and still pretty good for the DJF period. The 180-degree association then starts to break down (i.e. maxima and minima not occurring in opposite phases).

 

But FMA is really interesting to me, because there is a MAXIMUM, and a big max at that, in the EASTERLY phase of the QBO ... for phase -60, ZERO months below 33 percentile. This is followed by a rapid switch to a big minimum in late spring (AMJ) ... for phase -30, ZERO months above 67 percentile. That's pretty significant to me. And I'm really curious about that switch during the spring season.

This is more like it! lol I can't even believe you have time to do this kind of thing anyway with your schedule. What we don't want to do here is complicate the signal by getting more specific but I think the tri-monthly approach is fine. With all else being equal, the state of the QBO shear stress should have an effect on the arrival of any mid winter warmings and final warmings. The problem is: other factors like ENSO / solar etc. also have an effect on the waves and their timing. It seems like you were able to isolate the QBO's timings with this method but it is definitely early yet to say for sure.

The spring switch you are noting could definitely be related to the final warming. But this is also an interesting time of year, chemically-too, with the equinox/sun and what that may mean with the tropical and polar stratosphere (which presumably the QBO should have some effect on).

[HM]

There is some excellent symmetry there by the way! A good sign!

[Mike.Ventrice]

Honestly, I was working on this on the side of a lot of other stuff and didn't even think much further than I just wanted to check it out for the 4-month winter season. So you make an excellent point here. 

 

I just ran the same stats for every trimonthly period. I wouldn't want to try any less than three month periods because of sample size issues for each phase.

 

So for each of these periods (centered on the month listed), I calculated the 33% and 67% percentile thresholds (variable). I also used the full year percentile thresholds for comparison (constant).

 

Screen Shot 2013-04-24 at 1.52.40 PM.png

 

Screen Shot 2013-04-24 at 1.52.56 PM.png

 

So the strongest influence (with clear 180-degree association) appears in the NDJ period and still pretty good for the DJF period. The 180-degree association then starts to break down (i.e. maxima and minima not occurring in opposite phases).

 

But FMA is really interesting to me, because there is a MAXIMUM, and a big max at that, in the EASTERLY phase of the QBO ... for phase -60, ZERO months below 33 percentile. This is followed by a rapid switch to a big minimum in late spring (AMJ) ... for phase -30, ZERO months above 67 percentile. That's pretty significant to me. And I'm really curious about that switch during the spring season.

 

Had to sit back and read this more thoroughly.. I really like this and it's a great idea! I am not as educated as I believe I should be with regards to the qbo.. so I have some questions maybe you could help me with. In your movie of Velocity potential on your blog, why do you think the VP anomaly signature progresses eastward and is at times a wavenumber one feature, and other times a wave number 2? I presume the time transition of each phase for the qbo is slow, right? The way the VP signature looks its very similar to the MJO/CCKWs but the time scale has to be different... Just wondering  you could help me wrap my mind around this. Great work!

[OKpowdah]

Had to sit back and read this more thoroughly.. I really like this and it's a great idea! I am not as educated as I believe I should be with regards to the qbo.. so I have some questions maybe you could help me with. In your movie of Velocity potential on your blog, why do you think the VP anomaly signature progresses eastward and is at times a wavenumber one feature, and other times a wave number 2? I presume the time transition of each phase for the qbo is slow, right? The way the VP signature looks its very similar to the MJO/CCKWs but the time scale has to be different... Just wondering  you could help me wrap my mind around this. Great work!

 

I haven't investigated the relationship to velocity potential very much yet. I was trying to understand how the QBO might modulate the global pattern in longitudinal variation. So yeah, I was also trying to see how any patterns might project onto the MJO EOFs. Haven't revisited that yet, but plan on trying probably in a couple weeks.

 

The time transition for each phase is pretty slow, yes. Also, I'm splitting it into 30 degree increments because I felt like that was a convenient "resolution" to analyze the oscillation. I could do the same analysis at any "angle" ... 29 degrees, -36, 154.3. lol. But for the 30 degree increments, their average residence time is about 2.3 months.

 

Also this method uses a cosine-weight for each month. So take a month with a certain phase angle ... the months around it will have weights trailing off from 1 to 0. And that's what I use for all of these tests: a monthly array of weights for each phase.

[OKpowdah]

 

This is more like it! lol I can't even believe you have time to do this kind of thing anyway with your schedule. What we don't want to do here is complicate the signal by getting more specific but I think the tri-monthly approach is fine. With all else being equal, the state of the QBO shear stress should have an effect on the arrival of any mid winter warmings and final warmings. The problem is: other factors like ENSO / solar etc. also have an effect on the waves and their timing. It seems like you were able to isolate the QBO's timings with this method but it is definitely early yet to say for sure.

The spring switch you are noting could definitely be related to the final warming. But this is also an interesting time of year, chemically-too, with the equinox/sun and what that may mean with the tropical and polar stratosphere (which presumably the QBO should have some effect on).

 

 

I'm trying to figure out the best way of incorporating / filtering other factors like ENSO and solar. Any recommendations? I'm worried about how thinned out the data becomes if I keep binning it. Thinking about other methods..

[HM]

 

I'm trying to figure out the best way of incorporating / filtering other factors like ENSO and solar. Any recommendations? I'm worried about how thinned out the data becomes if I keep binning it. Thinking about other methods..

ENSO and solar make me cringe. The problems are the same in every paper. Either they filter ENSO out altogether which reduces the sample size or they try to simplify it with a ranking system. Both methods still do not properly take into account how ENSO affected the stratosphere and the time scale. I've read papers that did try to isolate ENSO's effect on polar warmings with some success.

Perhaps you could filter ENSO years that are anomalously strong out altogether first and see what you get (interestingly, no one has ever tried intraseasonal oscillation frequency as a metric for ENSO strength. It is always SST, MEI etc to help define it. I guess it doesn't matter for now). Maybe you can take your seasonal periods' percentiles and find what their average solar/ENSO conditions were. I'll definitely keep thinking about this and get back to you if I have any type of epiphany.  

[HM]

 

Maybe you can take your seasonal periods' percentiles and find what their average solar/ENSO conditions were. I'll definitely keep thinking about this and get back to you if I have any type of epiphany.

 

By this I mean for every case group into an AO ranking by its percentile. Each of these groups (the lowest to highest ranked AO) in each of the QBO phases could have a certain solar/enso configuration that explains the variance within that QBO phase. It's a long shot I know.

[aljareer]

Hi Guys

 

I think : Arctic Oscillation Index is not accurate enough. Because It measures the atmospheric layer 1000 hPa Only.

While index (NAM), it measures the most layers of the atmosphere.
It is more accurate than the index (AO)

http://cmdp.ncc.cma.gov.cn/Monitoring/cn_moni_stratosphere.php?Elem=NAMIndex&eYear=&eMonth=2&eDay=18&search=%E6%98%BE%E7%A4%BA%E5%9B%BE%E5%BD%A2

I watched the Arctic Oscillation Index over the ten years I think it is inaccurate.

The Arctic Oscillation is not a climatic phenomenon, it is a statistical indicator only, by : EOF analysis Method.

Thanks To All

[OKpowdah]

By the way, don't know why I didn't do this sooner, but I put together a height section for the MQI phases. This is plotting correlation to zonal wind by phase and pressure ... zonally averaged along the equator between 5S and 5N.

 

It's just taking 30-degree phase increments in the plot, so I can increase resolution of the graphic by creating additional time series for each phase at smaller increments. 

 

But for now, the result is very good. As should be expected, it matches the time-height sections of observed QBO cycles. It can serve as a good reference for what each phase represents.

[OKpowdah]

Same plot but for temperature

 

[OKpowdah]

One more ...

 

Temperature zonally averaged between 10N-20N ... "off-equator" and in the MDR latitude band.

 

[Ed Lizard]

One more ...

 

Temperature zonally averaged between 10N-20N ... "off-equator" and in the MDR latitude band.

 

tempofftimeheight.png

Totally missing something (me, not you)- at what longitude are those anomalies plotted versus phase and height at?  And does colder temps imply a higher tropopause?

 

Thanks.  I hope the question itself makes sense.

[OKpowdah]

Totally missing something (me, not you)- at what longitude are those anomalies plotted versus phase and height at?  And does colder temps imply a higher tropopause?

 

Thanks.  I hope the question itself makes sense.

 

Good questions. These plots are averaged across all longitudes. 

 

And yes! colder temperatures (seen by negative correlations) can imply a higher tropopause.

[mrgolf]

Hello guys! This msg is for either HM, mike ventrice, and/or whoever wants to try and answer it. I know its real early to be even discussing about this upcoming winter, but since we had a -qbo last winter, does that mean we will have a +qbo this coming winter? If so, do we want to see a higher solar max in tandem with that? I know we was on a record streak of a -qbo for the longest. I just wanted to know what you guys think? Lastly, if we see another enso neutral going through winter 2014, do you guys expect an earlier start to winter....meaning perhaps winter starts earlier this winter, but ends rather quick. Thanks kevin

[OKpowdah]

Very quick plot I just put together ... this is for DJF. MQI phase along the Y axis, longitude along the X axis. And a crude measure of the annular mode by z500 between 30-50N minus z500 between 60-90N ... so negative values are indicative of blocking.

 

Notice a dominant blocking regime across the entire hemisphere at phase 210 ... with the strongest signal in the western Pacific and the Atlantic.

 

 

[Mike.Ventrice]

Hello guys! This msg is for either HM, mike ventrice, and/or whoever wants to try and answer it. I know its real early to be even discussing about this upcoming winter, but since we had a -qbo last winter, does that mean we will have a +qbo this coming winter? If so, do we want to see a higher solar max in tandem with that? I know we was on a record streak of a -qbo for the longest. I just wanted to know what you guys think? Lastly, if we see another enso neutral going through winter 2014, do you guys expect an earlier start to winter....meaning perhaps winter starts earlier this winter, but ends rather quick. Thanks kevin

 

Sorry Kevin.. Not much of a seasonal prediction buff yet. HM or Adam may be in better off to answer this.

 

OK, interesting plot...  real quick, does the y axis represent phase where the x axis is... time?

 

Is your MQI running in real time yet?

 

Thanks!

[OKpowdah]

Sorry Kevin.. Not much of a seasonal prediction buff yet. HM or Adam may be in better off to answer this.

 

OK, interesting plot...  real quick, does the y axis represent phase where the x axis is... time?

 

Is your MQI running in real time yet?

 

Thanks!

 

Sorry I'm trying plotting a bunch of different things in limited time, which is why I haven't bothered even labeling axes ... great scientific practice right lol.

X axis is longitude, Y axis is phase.

 

The MQI is just based on monthly wind data, which the CPC has in real time, so I can definitely follow the current progression

[HM]

Very quick plot I just put together ... this is for DJF. MQI phase along the Y axis, longitude along the X axis. And a crude measure of the annular mode by z500 between 30-50N minus z500 between 60-90N ... so negative values are indicative of blocking.

 

Notice a dominant blocking regime across the entire hemisphere at phase 210 ... with the strongest signal in the western Pacific and the Atlantic.

 

So basically, once the westerly shear stress is completely gone, the potential for an Aleutian High and Greenland High increases. Notice in 2011-12, the QBO shear stress behaved much like that 180 to 210 period. The lingering westerly anomaly at the lowest levels likely aided in the +EPO/AO, especially in the first half of the winter.

[OKpowdah]

So basically, once the westerly shear stress is completely gone, the potential for an Aleutian High and Greenland High increases. Notice in 2011-12, the QBO shear stress behaved much like that 180 to 210 period. The lingering westerly anomaly at the lowest levels likely aided in the +EPO/AO, especially in the first half of the winter.

 

Definitely agree re: 11-12 winter ... also one of the reasons I think it's important I find some method of incorporating ENSO (and solar) ... The interaction of the descending easterly QBO regime on the Nina-induced pattern in the Pacific results in a pretty +EPO signal

[DaculaWeather]

I'm curious how this is looking at this point in the year? What made me think of it was DT's point about the QBO.