Let's talk ENSO

[40/70 Benchmark]

What is important is how the sst anomalies at a given posotion manifest themselves into the weather pattern around the hemisphere, not an arbitrary longitude where you feel as though said manifestation should take place.

Anyway, just my view....we are all different.

[40/70 Benchmark]

4 minutes ago, 40/70 Benchmark said:

I guess we are just going to have to agree to disagree on this. The anomaly in this composite is centered at like 165-170W, which is pretty far to the west, and that is reflected in the resultant H5 and VP (forcing) composites.

Its no coincidence why this composite gives us the warmest look...its west-based.

[weatherwiz]

6 hours ago, 40/70 Benchmark said:

 

You are referring to the intensity composites....yea, its because of the climo period. I used 1951-2010 for the temps, which will accentuate the positive anomalies due to the more expansive climo period encompassing a cooler overall global canvass several decades ago. However, the H5 composite would only allow me to utilize the 1981-2010 period, which attenuates the positive height anomalies because it excludes that same coolest stretch from several decades ago.

I really wish the H5 composite would allow me to use 1951-2010...I said this to Will the other day.

If you use this link you can customize the climo period

https://psl.noaa.gov/data/atmoswrit/map/index.html

Choose your source for the data set and then check subtract dataset 2 (for the dataset 2 box select the dataset)

in the Enter year range if different from dataset 1 (area highlighted in blue) input your range...this will be the climate period. 

For variable static make sure you click on mean. This will give the anomaly.

I emailed PSL well over a year ago and they gave me this link and told me these steps to make a customized climo period!

[It's Always Sunny]

9 hours ago, 40/70 Benchmark said:

Its no coincidence why this composite gives us the warmest look...its west-based.

Yeah I don't think there is a right or wrong in something like we said is pretty subjective and we are all for the most part in the same line of thought. Anyways, I updated the La Nina composites that I already had and comparing what I had for CP La Nina versus what you had for WP La Nina they're actually pretty similar despite the years not being identical, with warmer than average temps mainly across central-eastern US, which makes sense. If you jump onto my website though you won't see that close of a correlation  because I used the 1991-2020 climo which now has me contemplating changing it to maybe 1981-2010 and just keep it a sliding interval because my temp/precip anomalies are washed out because 1) I am using a -10-->10 contour interval w/ increments of 1 and 2) New climo raises ground zero so much that it's hard to discern anything unless it is "super anomalous" imo. Below are different samples of the same years, with my website image on the left, then the following two with different climos and intervals. Ahhh.

[40/70 Benchmark]

8 hours ago, weatherwiz said:

If you use this link you can customize the climo period

https://psl.noaa.gov/data/atmoswrit/map/index.html

Choose your source for the data set and then check subtract dataset 2 (for the dataset 2 box select the dataset)

in the Enter year range if different from dataset 1 (area highlighted in blue) input your range...this will be the climate period. 

For variable static make sure you click on mean. This will give the anomaly.

I emailed PSL well over a year ago and they gave me this link and told me these steps to make a customized climo period!

Thanks. I'll give that a shot. Like Summy said, the latest climo period does begin to wash out the anomalies, which defeats the purpose. That's why you need to use a larger climo period.

[weatherwiz]

9 hours ago, 40/70 Benchmark said:

Thanks. I'll give that a shot. Like Summy said, the latest climo period does begin to wash out the anomalies, which defeats the purpose. That's why you need to use a larger climo period.

Agreed...I too I think was originally using something like 1951-2010...but that too just didn't seem like enough to me, especially when doing analysis on years in the early 1900's or late 1800's. But as you know, data back then certainly has accuracy and validity questions and the majority of it I believe is all re-analysis and not actual data. 

One thing I've always wanted to do but I'm not good enough at math to do it (although I'm sure there are programs that can do this too but I can't program either  ) but when I start getting into NAO/AO I've always wanted to create a weekly/bi-weekly version of the product instead of just a monthly value. Now I know the height anomalies are significantly greater value than the raw number but you can use the raw numbers with the creation of the height anomalies. 

Daily data for the NAO/AO exist and you would think it would be as easy as just adding and dividing...but that doesn't work. For example, if you were to add the daily NAO values for say January 1950 and divide by 31...the answer does not equate to what is listed for the monthly value. I've tried to read how the monthly values are derived and EOF's but lie I said...that's way beyond my math skills. 

[40/70 Benchmark]

2 minutes ago, weatherwiz said:

Agreed...I too I think was originally using something like 1951-2010...but that too just didn't seem like enough to me, especially when doing analysis on years in the early 1900's or late 1800's. But as you know, data back then certainly has accuracy and validity questions and the majority of it I believe is all re-analysis and not actual data. 

One thing I've always wanted to do but I'm not good enough at math to do it (although I'm sure there are programs that can do this too but I can't program either  ) but when I start getting into NAO/AO I've always wanted to create a weekly/bi-weekly version of the product instead of just a monthly value. Now I know the height anomalies are significantly greater value than the raw number but you can use the raw numbers with the creation of the height anomalies. 

Daily data for the NAO/AO exist and you would think it would be as easy as just adding and dividing...but that doesn't work. For example, if you were to add the daily NAO values for say January 1950 and divide by 31...the answer does not equate to what is listed for the monthly value. I've tried to read how the monthly values are derived and EOF's but lie I said...that's way beyond my math skills. 

Well, I only use analogs back to 1950, so the 1951-2010 climo period is fine for me.

[CoastalWx]

I know they changed the classifications of ENSO events based on the new climo. For example some years that are strong Nina...are now considered moderate based on the new climo given warmer SSTs. However, has anyone ever tried to use where the forcing is as a better proxy? What I mean is that stronger Ninas have a basic H5 pattern based on where the forcing is in the Pacific. In theory, I don't think that take should change because the new ENSO classifications call a prior strong event, moderate based on a few tenths of a degree Celsius of warming. 

Lets say we have a neutral ENSO based on SST anomalies. However, 10 years ago this may have been considered weak Nina, and we currently have easterly wind anomalies as if it was a weak Nina. Something to think about. 

[weatherwiz]

11 minutes ago, CoastalWx said:

I know they changed the classifications of ENSO events based on the new climo. For example some years that are strong Nina...are now considered moderate based on the new climo given warmer SSTs. However, has anyone ever tried to use where the forcing is as a better proxy? What I mean is that stronger Ninas have a basic H5 pattern based on where the forcing is in the Pacific. In theory, I don't think that take should change because the new ENSO classifications call a prior strong event, moderate based on a few tenths of a degree Celsius of warming. 

Lets say we have a neutral ENSO based on SST anomalies. However, 10 years ago this may have been considered weak Nina, and we currently have easterly wind anomalies as if it was a weak Nina. Something to think about. 

That is a fantastic point. That is actually why when I did my composites I used various climo periods. If you were to look at the ONI which obviously dates back to 1950...it's comparing all those years to 1991-2020 climo...I mean is it really fair/accurate to compare an ENSO event in the 1950's or 1960's to the current climo? 

But my next step is to make the composites looking at tropical forcing (using OLR anomalies)...is that like what you're referring to with regarding to where the forcing is? I do agree that I think we would see stronger correlations looking at height patterns then we would with temperature anomalies. 

One thing to really consider too, especially regarding La Nina events is how the waters have warmed over the years. Average SST's are certainly warmer now than they were 50-60 years ago. Obviously the difference isn't substantial, but my point is is a weak La Nina now different from a weak La Nina from the 1950's or even earlier? We just classify ENSO based on temperature anomalies and longevity/consistency of these anomalies but what we consider a weak La Nina now may have been considered ENSO neutral...and with waters slightly warmer than they used to be this is certainly going to have some impact, especially with moisture content and latent heat. 

[CoastalWx]

46 minutes ago, weatherwiz said:

That is a fantastic point. That is actually why when I did my composites I used various climo periods. If you were to look at the ONI which obviously dates back to 1950...it's comparing all those years to 1991-2020 climo...I mean is it really fair/accurate to compare an ENSO event in the 1950's or 1960's to the current climo? 

But my next step is to make the composites looking at tropical forcing (using OLR anomalies)...is that like what you're referring to with regarding to where the forcing is? I do agree that I think we would see stronger correlations looking at height patterns then we would with temperature anomalies. 

One thing to really consider too, especially regarding La Nina events is how the waters have warmed over the years. Average SST's are certainly warmer now than they were 50-60 years ago. Obviously the difference isn't substantial, but my point is is a weak La Nina now different from a weak La Nina from the 1950's or even earlier? We just classify ENSO based on temperature anomalies and longevity/consistency of these anomalies but what we consider a weak La Nina now may have been considered ENSO neutral...and with waters slightly warmer than they used to be this is certainly going to have some impact, especially with moisture content and latent heat. 

Yes regarding your second paragraph.

 

But to your last paragraph...waters have warmed everywhere. I am just afraid that we may be missing some things due to the new classification.  Pretend we have a year where the waters near the dateline are -0.5C currently. The waters to the east off of Peru are near normal. 10 years ago maybe the waters would be near -0.9C near the dateline and -0.4C off Peru. However, is the atmosphere really behaving differently? We still have that temp differential...it's just that the temp baseline changed. 

I dunno...maybe I'm talking in circles, but it's why I asked regarding looking at the tropical forcing. 

[40/70 Benchmark]

1 hour ago, CoastalWx said:

I know they changed the classifications of ENSO events based on the new climo. For example some years that are strong Nina...are now considered moderate based on the new climo given warmer SSTs. However, has anyone ever tried to use where the forcing is as a better proxy? What I mean is that stronger Ninas have a basic H5 pattern based on where the forcing is in the Pacific. In theory, I don't think that take should change because the new ENSO classifications call a prior strong event, moderate based on a few tenths of a degree Celsius of warming. 

Lets say we have a neutral ENSO based on SST anomalies. However, 10 years ago this may have been considered weak Nina, and we currently have easterly wind anomalies as if it was a weak Nina. Something to think about. 

I included velocity potential along with the H5/temp and precip composites.

[CoastalWx]

8 minutes ago, 40/70 Benchmark said:

I included velocity potential along with the H5/temp and precip composites.

I'll take a look...admittedly have not sat down and looked at what you posted.

[40/70 Benchmark]

Just now, CoastalWx said:

I'll take a look...admittedly have not sat down and looked at what you posted.

If you look through the composites, it matches up pretty well with what you would expect., with descent above the coolest anomalies, and ascent aligned with the warmer.

[weatherwiz]

1 hour ago, CoastalWx said:

Yes regarding your second paragraph.

 

But to your last paragraph...waters have warmed everywhere. I am just afraid that we may be missing some things due to the new classification.  Pretend we have a year where the waters near the dateline are -0.5C currently. The waters to the east off of Peru are near normal. 10 years ago maybe the waters would be near -0.9C near the dateline and -0.4C off Peru. However, is the atmosphere really behaving differently? We still have that temp differential...it's just that the temp baseline changed. 

I dunno...maybe I'm talking in circles, but it's why I asked regarding looking at the tropical forcing. 

Gotcha...I see what you're saying and I share similar thoughts. But that is the big question...is the atmosphere behaving differently? My guess is that the atmosphere may behave a bit differently. 

Lets say the average water temperature within the ENSO region (3.4) was 27°C. So if you get 5-consecutive trimontly periods where that average is 26.3°C that's a pretty indicative La Nina episode. These colder waters are going to have some influence on convection, etc. 

Let's say now the average water temperature in the ENSO region is 28°C. So 5-consecutive trimonthly periods where the average is say 27.3°C would be a defined La Nina. So in theory, we have a La Nina given the anomalies with respect to the average, but these waters are still much warmer than they used to be and I would think this is going to yield a different atmospheric response than 26°C waters. 

So could there be a fair assumption (which would lead to the investigation) that weak La Nina's now may actually act more like how an ENSO neutral-warm or very weak EL Nino?

Rays composites regarding the tropical forcing certainly do shed some light into these ideas. Next week I'm going to plot OLR anomalies and do it with the climo breakdowns I did for temperature anomalies. While I haven't done anything with EL Nino yet I may also chart some weak EL Nino's and make a comparison as well.

[40/70 Benchmark]

28 minutes ago, weatherwiz said:

Gotcha...I see what you're saying and I share similar thoughts. But that is the big question...is the atmosphere behaving differently? My guess is that the atmosphere may behave a bit differently. 

Lets say the average water temperature within the ENSO region (3.4) was 27°C. So if you get 5-consecutive trimontly periods where that average is 26.3°C that's a pretty indicative La Nina episode. These colder waters are going to have some influence on convection, etc. 

Let's say now the average water temperature in the ENSO region is 28°C. So 5-consecutive trimonthly periods where the average is say 27.3°C would be a defined La Nina. So in theory, we have a La Nina given the anomalies with respect to the average, but these waters are still much warmer than they used to be and I would think this is going to yield a different atmospheric response than 26°C waters. 

So could there be a fair assumption (which would lead to the investigation) that weak La Nina's now may actually act more like how an ENSO neutral-warm or very weak EL Nino?

Rays composites regarding the tropical forcing certainly do shed some light into these ideas. Next week I'm going to plot OLR anomalies and do it with the climo breakdowns I did for temperature anomalies. While I haven't done anything with EL Nino yet I may also chart some weak EL Nino's and make a comparison as well.

I need to do en Nino composites, as well. I'll do those either next spring or summer.

[weatherwiz]

1 hour ago, 40/70 Benchmark said:

I need to do en Nino composites, as well. I'll do those either next spring or summer.

I only started with Nina given we're in a Nina state. I actually tried to start this project last year but I was really struggling with the climo period aspect because I didn't want to use 1981-2010 so I was going to go with 1951-2010 but I still wanted better. I'm hoping after this Nina breaks we'll get a few year period of ENSO neutral conditions...that way more work can be done with Nino composites and be ready for one  

[40/70 Benchmark]

2 minutes ago, weatherwiz said:

I only started with Nina given we're in a Nina state. I actually tried to start this project last year but I was really struggling with the climo period aspect because I didn't want to use 1981-2010 so I was going to go with 1951-2010 but I still wanted better. I'm hoping after this Nina breaks we'll get a few year period of ENSO neutral conditions...that way more work can be done with Nino composites and be ready for one  

Yea, obviously..same here. La nina was the priority. That changes next year. 

[ORH_wxman]

The borderline cases are all arbitrary anyway. There is no real fundamental difference between some year like 2008-2009 and 1967-1968 even though ‘08-09 is technically a weak Niña and ‘67-68 is negative neutral. 
 

Ironically about 10 years ago, ‘08-09 was considered negative neutral and ‘67-68 was a weak Niña before they adjusted everything. 

[weatherwiz]

One thing I want to add too is I was using two different datasets. I was using 20th Century Reanalysis V3 at first. This was because it covered data going back into the 1800's but it only goes to 2015. So when I did was use NCEP/NCAR R1 for years after 2015. However, when I did that I noticed some major discrepancies and this is a reported issue. For example, I think it was winter 2010-2011 which I believe was the winter with the big cold in the East (particularly into the mid-Atlantic region). The 20th century reanalysis did not reflect this. So I had to virtually start over and what I did was for years before 1981 I used 20th Century Reanalysis and after 1981 I used NCEP/NCAR as that's when that dataset began.  

[weatherwiz]

1 minute ago, ORH_wxman said:

The borderline cases are all arbitrary anyway. There is no real fundamental difference between some year like 2008-2009 and 1967-1968 even though ‘08-09 is technically a weak Niña and ‘67-68 is negative neutral. 
 

Ironically about 10 years ago, ‘08-09 was considered negative neutral and ‘67-68 was a weak Niña before they adjusted everything. 

Yes, this is a great point. Another year that was exceptionally borderline was '64-'65. Ultimately you would probably want to focus on the stronger type events and draw up composites that way and then when you have smaller events you can make a decision as to how much influence that weak event may really have. 

I suppose another thing to consider is when an event started/ended. There are a handful of Nina's that officially began during the fall months or even during the winter. Those may yield a different output than years when the event was already established. 

[weatherwiz]

Well here's a start. Will build more on this next week doing OLR anomalies by strength, structure (I'll be combining central/basin-wide), etc. 

This is for all La Nina's and using the ensemble ONI matric

I'm also wondering with these composites if it would be even better to use a consistent legend. 

[40/70 Benchmark]

On 9/1/2022 at 8:21 AM, It's Always Sunny said:

I have several instances where there is a strong emphasis west of 150 but what Paul has in his is west of 170 which I had one borderline instance but nothing else. West of 150 with not much more west of 170W is still a CP La Niña to me.

A good rule of thumb is that if you can't find a single season over like a 100 year sample size worth of data that meets your criteria, then it may be too exclusive, but to each their own.  Maybe you could at least split your "central based" composite into a couple of subsets because there is most certain a difference between an eastward leaning central based season, like 2010-2011, and a season that I consider to be west based, like 2011-2012. That needs to be reflected IMHO.

[40/70 Benchmark]

5 hours ago, ORH_wxman said:

The borderline cases are all arbitrary anyway. There is no real fundamental difference between some year like 2008-2009 and 1967-1968 even though ‘08-09 is technically a weak Niña and ‘67-68 is negative neutral. 
 

Ironically about 10 years ago, ‘08-09 was considered negative neutral and ‘67-68 was a weak Niña before they adjusted everything. 

I always considered '08-'09 as weak la nina.

[ORH_wxman]

10 hours ago, 40/70 Benchmark said:

I always considered '08-'09 as weak la nina.

It certainly acted like a weak Niña with a deeply negative PDO and -PNA

[It's Always Sunny]

 

On 9/2/2022 at 4:01 PM, weatherwiz said:

Well here's a start. Will build more on this next week doing OLR anomalies by strength, structure (I'll be combining central/basin-wide), etc. 

This is for all La Nina's and using the ensemble ONI matric

I'm also wondering with these composites if it would be even better to use a consistent legend. 

That's how I approached my composites regarding SSTa, 500 gph anoms & surface temp anoms. Only exception was precip anomalies because it was far too variable so I allowed that to be a sliding interval. Using a consistent interval allows not only easier cross-analysis but also helps determine stronger vs weaker events/anomalies because everything has an equal baseline. 

 

On 9/2/2022 at 8:06 PM, 40/70 Benchmark said:

A good rule of thumb is that if you can't find a single season over like a 100 year sample size worth of data that meets your criteria, then it may be too exclusive, but to each their own.  Maybe you could at least split your "central based" composite into a couple of subsets because there is most certain a difference between an eastward leaning central based season, like 2010-2011, and a season that I consider to be west based, like 2011-2012. That needs to be reflected IMHO.

I have Nov-Mar 2010-2011 under basin wide but regardless, I understand what you're saying; since I'm having a hard time calling something "west based", a good approach could be to call them "east leaning" or "west leaning". I like it. Once I sort through more composites again I'll probably start binning them that way as my sample size grows.

 

 

Regarding the disco of using tropical forcing as a proxy for ENSO: While I think that's possible and a really good idea (and could be better than SSTa imo), I just feel like there are too many external variables that can impact velocity potential (or even OLR) anomalies beyond what is occurring in the Pacific Ocean. Also what anomaly period do you use for that which is another question.  What about mean 500mb gph because then that way you don't need to use a climo period and not have to worry about SST changing. I think upper air patterns could be easily correlated with ENSO episodes but to discern what may be strong, weak or neutral could be tricky. Haven't tried it so idk but that's where my mind is heading. I've spent the past few days thinking about what climo period to use for my composites and the more I thought about it I think Paul's method makes the most sense because it classifies something "for that period in history" because it's not scientifically sound to use a current day baseline for an event that happened 60 years ago. Food for thought but maybe it's not scientifically sound to compare an ENSO event in the 1960s to one within the last decade due to the warming atmosphere (water and land temps). Not saying I agree or disagree with that but something I thought of. I like the idea of using 1950-2010 b/c it encompasses everything but 1950 vs. 2010 are also two very different time periods. Again I don't think there is a right or wrong but many subjective ways of doing it. 

[It's Always Sunny]

On 9/2/2022 at 8:08 PM, 40/70 Benchmark said:

I always considered '08-'09 as weak la nina.

 

On 9/3/2022 at 6:19 AM, ORH_wxman said:

It certainly acted like a weak Niña with a deeply negative PDO and -PNA

I have that down as a weak La Nina as well.

[40/70 Benchmark]

@It's Always SunnyI have 2010-2011 as basin wide, as well, but it has an east-tilt...similar to last year, though least year ended up flat out east-based in my book. My point is that if you are going to have the basin wide data set that expansive, its important to have some type of delineation... we agree on that. I consider 2011-2012 west based, and while you can technically call it basin wide, it's a different animal than 2010-2011.

[weatherwiz]

13 hours ago, It's Always Sunny said:

 

That's how I approached my composites regarding SSTa, 500 gph anoms & surface temp anoms. Only exception was precip anomalies because it was far too variable so I allowed that to be a sliding interval. Using a consistent interval allows not only easier cross-analysis but also helps determine stronger vs weaker events/anomalies because everything has an equal baseline. 

 

I have Nov-Mar 2010-2011 under basin wide but regardless, I understand what you're saying; since I'm having a hard time calling something "west based", a good approach could be to call them "east leaning" or "west leaning". I like it. Once I sort through more composites again I'll probably start binning them that way as my sample size grows.

 

 

Regarding the disco of using tropical forcing as a proxy for ENSO: While I think that's possible and a really good idea (and could be better than SSTa imo), I just feel like there are too many external variables that can impact velocity potential (or even OLR) anomalies beyond what is occurring in the Pacific Ocean. Also what anomaly period do you use for that which is another question.  What about mean 500mb gph because then that way you don't need to use a climo period and not have to worry about SST changing. I think upper air patterns could be easily correlated with ENSO episodes but to discern what may be strong, weak or neutral could be tricky. Haven't tried it so idk but that's where my mind is heading. I've spent the past few days thinking about what climo period to use for my composites and the more I thought about it I think Paul's method makes the most sense because it classifies something "for that period in history" because it's not scientifically sound to use a current day baseline for an event that happened 60 years ago. Food for thought but maybe it's not scientifically sound to compare an ENSO event in the 1960s to one within the last decade due to the warming atmosphere (water and land temps). Not saying I agree or disagree with that but something I thought of. I like the idea of using 1950-2010 b/c it encompasses everything but 1950 vs. 2010 are also two very different time periods. Again I don't think there is a right or wrong but many subjective ways of doing it. 

The more I think about it I think that's the way I'm going to go as well. I'm going to take some time and re-do what I have saved in terms of maps/printouts and try and use a consistent legend. I've also gone back and utilized Ray's method of combining basin-wide and central-based into one (basin-wide). There are a few events though (just like 70-71 ) where they may not be a true basin wide with the anomalies shifted farther east but I think for something like that you just make a mental note. 

Interesting analysis regarding the use of tropical forcing. I had started making composites with this and some initial look would yield the notion of external forces acting. Another challenge too is you can have scenarios where there are multiple regions of increased convection. Of course too La Nina's can be a bit more difficult with the use of convection and tropical forcing given the colder waters typically yield more subsidence. So using EL Nino events may be better to build some background/understand regarding tropical forcing and influences on upper pattern.

[weatherwiz]

Also given the changes I have done/am doing at some point I'm going to delete all the original graphics/GIFs I posted and re-upload them. There was a few ENSO events I forgot to add into the structure category. 

[40/70 Benchmark]

On 9/1/2022 at 2:49 PM, weatherwiz said:

If you use this link you can customize the climo period

https://psl.noaa.gov/data/atmoswrit/map/index.html

Choose your source for the data set and then check subtract dataset 2 (for the dataset 2 box select the dataset)

in the Enter year range if different from dataset 1 (area highlighted in blue) input your range...this will be the climate period. 

For variable static make sure you click on mean. This will give the anomaly.

I emailed PSL well over a year ago and they gave me this link and told me these steps to make a customized climo period!

Man, I really hate the new color scheme it uses.....so that uses the 1951-2010 period because it was subtracted, even though it still says 1991-2020 at the top?