Let's talk ENSO

[weatherwiz]

3 minutes ago, 40/70 Benchmark said:

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?

I'm not a fan of the background or color scheme either. 

hmmm this is what I got using your parameters

 

Here is the input fields. Maybe you forgot to fill both boxes in with NCEP/NCAR?

 

[40/70 Benchmark]

42 minutes ago, weatherwiz said:

I'm not a fan of the background or color scheme either. 

hmmm this is what I got using your parameters

 

Here is the input fields. Maybe you forgot to fill both boxes in with NCEP/NCAR?

 

[40/70 Benchmark]

You can change the color scheme, as well:

[weatherwiz]

24 minutes ago, 40/70 Benchmark said:

You can change the color scheme, as well:

Sweet...didn't know you could change the color scheme. 

Ahhh I think I see why yours is saying 1990-2020 climo. Make sure you have mean checked off and not anomaly in variable statistic. Even though it says mean you'll get the map as an anomaly as its subtracting the difference from your input year against your desired climo period.

[It's Always Sunny]

3 hours ago, weatherwiz said:

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. 

Yeah hopefully today I am going to redo my temp anomaly maps and refine those a bit more to make some anomalies "pop". I think your approach regarding your selection of climo method is the best but I'm going to keep the 1991-2020 for mine just b/c everything else on my site has that at the moment. I'd also like to add more years to my dataset as well just to have a larger sample size of things. Good thread Paul thanks for starting this up.

[weatherwiz]

18 hours ago, It's Always Sunny said:

Yeah hopefully today I am going to redo my temp anomaly maps and refine those a bit more to make some anomalies "pop". I think your approach regarding your selection of climo method is the best but I'm going to keep the 1991-2020 for mine just b/c everything else on my site has that at the moment. I'd also like to add more years to my dataset as well just to have a larger sample size of things. Good thread Paul thanks for starting this up.

I've been wanting to start something like this for a while! It's great to bounce ideas and thoughts back and forth. I've had a vested interest in this area for like 15-years now and I've always struggled to really get going with research and analysis...mostly because I have a tendency to overanalyze and get too granular (hence the separation of basin wide vs. central based). 

I don't think its a bad idea either to compare to the most recent climatological period. I think there is some significant value in this because you can compare earlier years to an earlier climate period to the most recent climo period...doing this I think you can spot trends and see perhaps what areas are being influenced more heavily by recent warming trends.  

[It's Always Sunny]

On 9/6/2022 at 3:20 PM, It's Always Sunny said:

Yeah hopefully today I am going to redo my temp anomaly maps and refine those a bit more to make some anomalies "pop". I think your approach regarding your selection of climo method is the best but I'm going to keep the 1991-2020 for mine just b/c everything else on my site has that at the moment. I'd also like to add more years to my dataset as well just to have a larger sample size of things. Good thread Paul thanks for starting this up.

I went through everything again and added a few more years to my dataset however I elected to keep the temp anomaly maps the same because I wasn't seeing much of a difference using 0.5 intervals vs. 1.0. Next step is to sort through weak vs strong events and see what I can pull from that.

[weatherwiz]

12 minutes ago, It's Always Sunny said:

I went through everything again and added a few more years to my dataset however I elected to keep the temp anomaly maps the same because I wasn't seeing much of a difference using 0.5 intervals vs. 1.0. Next step is to sort through weak vs strong events and see what I can pull from that.

I was struggling earlier with the intervals as well when re-doing temp maps. I used a range of -7 to 7 and interval of 1. I did it for All La Nina winters and weak La Nina winters (using the Ensemble ONI) and just doing a consistent legend yielded a much better look for re-analysis! 

[It's Always Sunny]

I finished classifying my ENSO phases as Weak, Strong or Very Strong by taking the average ONI over the NDJ, DJF, JFM periods rather than the peak. I also included trends (Strengthening, Steady or Weakening) during those three overlapping seasons since I think this is an unexplored area of study. Nothing came to light as I was doing it but I also haven't cross-analyzed them yet either. FWIW I didn't have any Strengthening episodes which I was actually a bit surprised by, especially during La Nina episodes. My composites do not line up perfectly with the listed ENSO phase due to the climo period I used, however I stumbled across this link that explains CPC's strategy of using a continually updated 30-year base period for ONI every 5 years to prevent the warming trend which has great logic. The ONI I used already has this new 30-year base period methodology applied. I may revisit my composites and experiment applying the method the CPC used but it's going to take a while so I'm in no rush to do it atm. 

https://www.cpc.ncep.noaa.gov/products/analysis_monitoring/ensostuff/ONI_change.shtml

[40/70 Benchmark]

On 9/6/2022 at 2:07 PM, weatherwiz said:

I'm not a fan of the background or color scheme either. 

hmmm this is what I got using your parameters

 

Here is the input fields. Maybe you forgot to fill both boxes in with NCEP/NCAR?

 

Paul, I don't see a velocity potential option on this like the other one.....I hate that, as I use it to show ENSO forcing. I guess OLR would be the closest proxy to that of the available option with this tool.... @CoastalWxis OLR pretty much akin to .2101 sigma VP?

[weatherwiz]

53 minutes ago, 40/70 Benchmark said:

Paul, I don't see a velocity potential option on this like the other one.....I hate that, as I use it to show ENSO forcing. I guess OLR would be the closest proxy to that of the available option with this tool.... @CoastalWxis OLR pretty much akin to .2101 sigma VP?

Great question! 

[weatherwiz]

15 hours ago, It's Always Sunny said:

I finished classifying my ENSO phases as Weak, Strong or Very Strong by taking the average ONI over the NDJ, DJF, JFM periods rather than the peak. I also included trends (Strengthening, Steady or Weakening) during those three overlapping seasons since I think this is an unexplored area of study. Nothing came to light as I was doing it but I also haven't cross-analyzed them yet either. FWIW I didn't have any Strengthening episodes which I was actually a bit surprised by, especially during La Nina episodes. My composites do not line up perfectly with the listed ENSO phase due to the climo period I used, however I stumbled across this link that explains CPC's strategy of using a continually updated 30-year base period for ONI every 5 years to prevent the warming trend which has great logic. The ONI I used already has this new 30-year base period methodology applied. I may revisit my composites and experiment applying the method the CPC used but it's going to take a while so I'm in no rush to do it atm. 

https://www.cpc.ncep.noaa.gov/products/analysis_monitoring/ensostuff/ONI_change.shtml

Great idea classifying that way! 

while I didn’t include this in my initial notes or within any work I have taken similar notes as well regarding trends. 

[CoastalWx]

1 hour ago, 40/70 Benchmark said:

Paul, I don't see a velocity potential option on this like the other one.....I hate that, as I use it to show ENSO forcing. I guess OLR would be the closest proxy to that of the available option with this tool.... @CoastalWxis OLR pretty much akin to .2101 sigma VP?

It is a decent proxy, yes. 

[40/70 Benchmark]

9 minutes ago, CoastalWx said:

It is a decent proxy, yes. 

Seems to line up pretty well with VP maps.

[40/70 Benchmark]

I have updated all of my la nina H5 and SST composites to reflect the 1951-2010 climo period. I ended up leaving the VP composites alone, which are all 1981-2010, save for east based (1991-2020). I had to update the east-based composite to include last season, so that meant 1981-2010 was no longer an option. I like how the VP reflects forcing a bit better than OLR, and VP is not an option of the climo period converter maps, unfortunately. I don't think heterogeneous climo periods for something like forcing is a huge deal.

[weatherwiz]

20 minutes ago, 40/70 Benchmark said:

I have updated all of my la nina H5 and SST composites to reflect the 1951-2010 climo period. I ended up leaving the VP composites alone, which are all 1981-2010, save for east based (1991-2020). I had it to update to include last season, so that meant 1981-2010 was no longer an option. I like how the VP reflects forcing a bit better than OLR, and VP is not an option of the climo period converter maps, unfortunately. I don't think heterogeneous climo periods for something like forcing is a huge deal.

That's a great point. 

I'm hoping to re-upload the GIFs in the original post sometime over the weekend. 

[40/70 Benchmark]

9 minutes ago, weatherwiz said:

That's a great point. 

I'm hoping to re-upload the GIFs in the original post sometime over the weekend. 

The main reason we even look at forcing is to illustrate a how disparate types of ENSO manifest around the hemisphere, so I really don't see the benefit of using a climo period and/or variable that attenuates said difference. It defeats the purpose. If the forcing was actually changing, then that is one thing...but we know from the H5 plots that it is not.

[weatherwiz]

On 9/8/2022 at 12:14 PM, CoastalWx said:

It is a decent proxy, yes. 

what are your thoughts regarding average anomaly vs. standardized anomaly. 

[40/70 Benchmark]

1 hour ago, weatherwiz said:

what are your thoughts regarding average anomaly vs. standardized anomaly. 

Or temp anomalies vs daily max temp anomalies...

[weatherwiz]

4 minutes ago, 40/70 Benchmark said:

Or temp anomalies vs daily max temp anomalies...

Temp is averaged daily temp (max + min / 2) 

[40/70 Benchmark]

1 minute ago, weatherwiz said:

Temp is averaged daily temp (max + min / 2) 

Right...just saying, is it better to use temps, or daily maxes...I have traditionally used temps...

[40/70 Benchmark]

Max would drown out the effect of climate change a bit...good alternative if you dnt want to use the 1951-2010 climo set.

[weatherwiz]

1 hour ago, 40/70 Benchmark said:

Right...just saying, is it better to use temps, or daily maxes...I have traditionally used temps...

Ohh I see what you’re saying. 

[It's Always Sunny]

1 hour ago, 40/70 Benchmark said:

Max would drown out the effect of climate change a bit...good alternative if you dnt want to use the 1951-2010 climo set.

Agreed, the daily average gives the dataset a larger sample size to work with.

[weatherwiz]

2 hours ago, 40/70 Benchmark said:

Max would drown out the effect of climate change a bit...good alternative if you dnt want to use the 1951-2010 climo set.

I should add…this wasn’t necessarily in reference to the ENSO work. 
 

I was working on something for work comparing summer of 2022 vs. 2021. I was looking at mean daily max, daily min, and mean avg temp along with anomalies. 
 

I’ve always used standardized anomalies in the past (for no other reason than it sounded cool lol) but for hahas wanted to see the difference between anomalies and standardized…and well that is a super large difference. 

[weatherwiz]

Ahh I did some digging on the site https://psl.noaa.gov/data/usclimdivs/help/ and found this information. 

I think when choosing standardized anomaly you're getting the anomalies in standard deviation (I guess that should have been obvious). So In the examples I posted the one on the right is -3SD to +3SD. for some reason though I guess they don't show the unit of measurement in the legend anymore. In the example on the link they have °F and SD in the legend. 

[40/70 Benchmark]

On 9/6/2022 at 2:07 PM, weatherwiz said:

I'm not a fan of the background or color scheme either. 

hmmm this is what I got using your parameters

 

Here is the input fields. Maybe you forgot to fill both boxes in with NCEP/NCAR?

 

I am having issues with that site all of a sudden. When I try to subtract 1951-2010 period, its defaulting to 1991-2020...

[weatherwiz]

1 hour ago, 40/70 Benchmark said:

I am having issues with that site all of a sudden. When I try to subtract 1951-2010 period, its defaulting to 1991-2020...

hmmm I should get some time to do some more stuff with this in a bit and I'll see if I have the same issues. 

I'm hoping to re-upload all the GIFs with the corrections today too. 

[weatherwiz]

2 hours ago, 40/70 Benchmark said:

I am having issues with that site all of a sudden. When I try to subtract 1951-2010 period, its defaulting to 1991-2020...

It's working for me. 

Make sure in the variable statistic you have mean selected and not anomaly and your dataset 1 and dataset 2 options are the same. 

[40/70 Benchmark]

3 minutes ago, weatherwiz said:

It's working for me. 

Make sure in the variable statistic you have mean selected and not anomaly and your dataset 1 and dataset 2 options are the same. 

OMFG, I kept forgetting to select the second data set after checking the group 2 box...my god.

Getting old.