ENSO Temps and Observations

[Michael Snyder]

Yea, it's another thread about El Nino- But this time I've got some numbers to back things up.

There are several ways you can measure the correlation between ENSO temps/events.

For now I have used this method:
I compared tornado data by using the ENSO conditions/temps the Dec/Jan PRIOR to the Tornado data.
(Not necessarily full blown El Nino's/La Nina's)

So what did each of the reports show?
Prior Dec/Jan ENSO conditions with average tornadoes per months of March/April 1985-2014
NE- 12.8 La Nina / 9.0 Neutral / 4.3 El Nino : La Nina kicks off an earlier Tornado season up north
CO- 6.8 El Nino / 1.1 Neutral / 2.6 La Nina: A win for El Nino, although March slightly favors La Nina in CO
KS- 21.8 La Nina / 17.6 Neutral / 14.2 El Nino : La Nina strong signal for early start in KS
OK- 28.8 La Nina / 16.6 Neutral / 10.8 El Nino : La Nina again with another very strong signal in OK for Mar/Apr
TX- 39.8 Neutral / 38.7 La Nina / 32.4 El Nino : Chalk one up for neutral years, with LA Nina>El Nino

There is a very strong signal for La Nina years especially in OK, KS, NE
-----------------------------------------------------------------------------------------------------------
Lets try May..........
Prior January ENSO conditions with average tornadoes per months of May 1985-2014
NE- 21.6 Neutral / 14.1 El Nino / 11.1 La Nina
CO- 10.0 El Nino / 9.2 Neutral / 9.5 La Nina
KS- 34.9 Neutral / 30.2 La Nina / 29.6 El Nino
OK- 28.4 El Nino / 26.4 La Nina / 18.6 Neutral
TX- 52.8 Neutral / 37.8 La Nina / 36.8 El Nino

The underlying message here is that the effects of La Nina seems to be out the window by May in most cases, with a maximum in neutral years
--------------------------------------
A little more clarification:
A reason for taking the previous winter ENSO status data is because ENSO temps are always on the decrease or the increase as you move into the spring months. A strong El Nino is not going to reflect those temps as you get into spring and the opposite is true for La Nina.
For example: For the period Mar-July 1985-2014 the Lowest ENSO temp is: -1.3, and the warmest ENSO is 1.4

So in other words you are never going to be able to show differences that can include Strong ENSO events and tornado data just by taking the raw ENSO temps and plugging in Tornado data. It gets skewed quickly since during the months of large variations of ENSO temps, the tornado activity is usually at a minimum (Oct-Feb, the same months which ENSO temps reach their extremes).

For those of you who want to see those numbers, Ill do those next. For any other requests, let me know.

I'm going to break this down into tornado rating also in the near future.

If anyone wants the raw data and spreadsheet I'm using,just let me know.

 

[Jeff Duda]

Good work!

I will say that statistical variations associated with low sample size probably pervade the data. In other words, since the observational record only has about 10 or so non-neutral events each, it's unlikely that many of these differences are statistically significant, particularly when the differences in tornado count between the ENSO classifications for a given state are only in the range of 5.0 and less.

In general your results suggest not to expect any increased tornado activity this spring (coming out of an El Nino), and perhaps some reduced numbers. If it remains very dry across the southern/central US through the rest of the winter and into spring, I would tend to think reduced tornado activity may be likely. However, given how ridiculously wet 2015 was across most of the US, with most of that deeper soil moisture remaining stored away as vegetation isn't pulling much of it up right now, I'm inclined to suspect tornado activity will be closer to normal this year rather than below normal.

Have you looked at June and July numbers? I suspect the signal will be weaker still, but you never know.

 

[Michael Snyder]

Good work!

I will say that statistical variations associated with low sample size probably pervade the data. In other words, since the observational record only has about 10 or so non-neutral events each, it's unlikely that many of these differences are statistically significant, particularly when the differences in tornado count between the ENSO classifications for a given state are only in the range of 5.0 and less.

In general your results suggest not to expect any increased tornado activity this spring (coming out of an El Nino), and perhaps some reduced numbers. If it remains very dry across the southern/central US through the rest of the winter and into spring, I would tend to think reduced tornado activity may be likely. However, given how ridiculously wet 2015 was across most of the US, with most of that deeper soil moisture remaining stored away as vegetation isn't pulling much of it up right now, I'm inclined to suspect tornado activity will be closer to normal this year rather than below normal.

Have you looked at June and July numbers? I suspect the signal will be weaker still, but you never know.

Exactly right Jeff, there are 12 La Nina / 9 Neutral / 9 El Nino periods during these 30 years. The Sample size sucks, and if we try to go back farther , then we run into tornado data issues such as underreported events compared to what gets reported now. Something to look forward to I guess, better and more consistent tornado reports.

Ok, so here is June:
NE- 22.0 LA / 20.1 N / 17.4 EL
CO- 22.1 N / 19.3 EL / 13.2 LA
KS- 22.4 EL / 20.4 LA / 14.2 N
OK- 11.5 LA / 9.1 EL / 4.7 N
TX- 33.1 EL / 22.0 N / 15.3 LA

After the 1998 Power El Nino, Kansas had a robust 41 Tornadoes that next June
And even with the TX huge El Nino Bias in June, TX only had 6 in the June of 1998
OK also had an above average Post 1998 El Nino June with 23, same with NE at 28 in June of 1998.

Another interesting thing is the lack of Tornadoes for OK AFTER a Neutral winter for May-June.

 

[Michael Snyder]

Good work!

I will say that statistical variations associated with low sample size probably pervade the data. In other words, since the observational record only has about 10 or so non-neutral events each, it's unlikely that many of these differences are statistically significant, particularly when the differences in tornado count between the ENSO classifications for a given state are only in the range of 5.0 and less.

In general your results suggest not to expect any increased tornado activity this spring (coming out of an El Nino), and perhaps some reduced numbers. If it remains very dry across the southern/central US through the rest of the winter and into spring, I would tend to think reduced tornado activity may be likely. However, given how ridiculously wet 2015 was across most of the US, with most of that deeper soil moisture remaining stored away as vegetation isn't pulling much of it up right now, I'm inclined to suspect tornado activity will be closer to normal this year rather than below normal.

Have you looked at June and July numbers? I suspect the signal will be weaker still, but you never know.

So yes, the La Nina signal was weaker still in July, but that high El Nino average for Texas in July is interesting.

 

[samuel stone]

Yeah it really seems difficult to draw any major conclusions from the may and june data. El nino up, El nino down, La nina, up la nina down neutral all over, with no real firm multi-state/regional geographic biases to take from it either. (for may and june at least). And yes I'm in complete agreement with the sample size issues. I've always preached about the difficulty of drawing any significant conclusions about tornado climatology (other than the obvious ones like may gets the most) because our data set of tornadoes is not only short in life (given the relative rarity of the phenomenon), but also prone to many biases (e.g. historically fewer people to report them in the western high plains/possibly harder to notice night time qlcs tornadoes in the east.. and on and on). These sort of issues have improved a lot but really only as of lately.

Gotta say though, these statistics really are fun as hell.. Keep em comin.

 

[Michael Snyder]

Here is the total for all 5 states (NE,CO,KS,OK,TX) Mar-July 1985-2014
With actual ENSO temps during Mar-July months ( previous Winter ENSO conditions ARE NOT calculated here)
La Nina -1.4 to -0.5 35 total Months
Neutral -0.4 to 0.4 87 total Months
El Nino 0.5 to 1.3 28 total Months

Mar-July Average
La Nina 51.6
Neutral 69.2
El Nino 59.6

Bring on the Neutral Years! El Nino gives some hope!

NOW, Let's change the numbers a tad to give the El Nino and La Nina years a few more months and see what happens. To do this I'll calculate with the following parameters:
This is what I got...
La Nina -1.4 to -0.3 51 total Months
Neutral -0.2 to 0.2 52 total Months
El Nino 0.3 to 1.4 47 total Months

Mar-July Average
La Nina 56.3
Neutral 70.5
El Nino 63.1

Again with the Neutral Years!
Side note- It lines up with the Theory that the PNW gets its strongest Mid latitude cyclones during Ocean transition years(months??) also.

And again with the signal that El Nino is not all bad.

The latest IRI/CPC Pacific Nino 3.4 SST has the Pacific moving to sub 0.5 ENSO temps sometime in May and down to 0.2 by June.


Ill go into more depth state by state and month by month here in the next week.

 

[calvinkaskey]

Last year the states right on the east coast were almost 100 tornadoes below average.

 

[Randy Jennings]

If it remains very dry across the southern/central US through the rest of the winter and into spring, I would tend to think reduced tornado activity may be likely.

NWS FWD agrees with you on this Jeff. They have a nice writeup on "A Dry El Nino" at http://www.srh.noaa.gov/fwd/?n=climateoutlook.

There have been several papers written on this subject. Here are 3 of them:

http://www.spc.noaa.gov/publications/schaefer/el_nino.htm
http://www.spc.noaa.gov/publications/cook/enso-mwr.pdf
http://www.spc.noaa.gov/publications/schaefer/sndgparm.pdf

My favorite quote from these papers:

They conclude that tornado occurrences do not favor one ENSO phase but rather exhibit a shift in geographic location.

 

[Kyle Brittain]

@Randy Jennings indeed, and so far this winter's events seem to confirm the theory of the Cook/Schaefer paper - that there are observed differences in tornado occurrences, especially in the southern states where the ENSO signal is stronger between EN/N/LN seasons. Most tornado activity, unsurprisingly, occurs near the typical position of the jet depending on ENSO phase: winter tornadoes in EN are typically confined to the gulf coast; they extend northeastward along the Mississippi and Ohio river valleys in LN winters; and they can occur anywhere from the plains to the Great Lakes to the mid-Atlantic and south in neutral winters.

This winter's pattern so far seems quite typical then:

 

[Kyle Brittain]

Most models are predicting a shift from the still currently strong El Nino to neutral by June, and many to a possible La Nina in fall:



Just to confirm though, as it's still a little unclear to me...do you go with the previous winter's ENSO status, or the current status (if in say, May/June) when considering tornado patterns, or the numbers from posts above? I could see the previous winter's status playing a pivotal role if it determines the amount of moisture that is available to storm formation in summer (ie here in Alberta where we seldom see gulf moisture, the amount of ground water from precip events (or lack thereof) determines moisture availability in summer). However, I could also see the typical position of the jet/storm track from the current time (say it transitions to neutral, like it may by this summer) being an equally important factor. During years like this, when we will likely be transitioning out of a strong El Nino into a neutral phase during the spring months, it seems like there will be a lot of complicating factors determining what we can expect. So would we expect tornado numbers to look more like the EN ones in April from the past average, and the neutral ones in June? Or do would we expect the whole season to be somewhat closer overall purely to an average EN season?

As an aside, I also found this article helpful:

http://iri.columbia.edu/news/enso-tornado/

(Keep in mind it's from a year ago, so it contains a prediction for last year's season, entering into EN)

 

[Michael Snyder]

Most models are predicting a shift from the still currently strong El Nino to neutral by June, and many to a possible La Nina in fall:

View attachment 11731

Just to confirm though, as it's still a little unclear to me...do you go with the previous winter's ENSO status, or the current status (if in say, May/June) when considering tornado patterns, or the numbers from posts above? I could see the previous winter's status playing a pivotal role if it determines the amount of moisture that is available to storm formation in summer (ie here in Alberta where we seldom see gulf moisture, the amount of ground water from precip events (or lack thereof) determines moisture availability in summer). However, I could also see the typical position of the jet/storm track from the current time (say it transitions to neutral, like it may by this summer) being an equally important factor. During years like this, when we will likely be transitioning out of a strong El Nino into a neutral phase during the spring months, it seems like there will be a lot of complicating factors determining what we can expect. So would we expect tornado numbers to look more like the EN ones in April from the past average, and the neutral ones in June? Or do would we expect the whole season to be somewhat closer overall purely to an average EN season?

As an aside, I also found this article helpful:

http://iri.columbia.edu/news/enso-tornado/

(Keep in mind it's from a year ago, so it contains a prediction for last year's season, entering into EN)

Ok, nevermind I see what you are saying here Kyle.

 

[Jeff Duda]

Just to confirm though, as it's still a little unclear to me...do you go with the previous winter's ENSO status, or the current status (if in say, May/June) when considering tornado patterns, or the numbers from posts above?

People typically consider ENSO conditions during the cold season prior to a given spring/summer season, for the following reasons:
1) For an attempt at predictability
2) ENSO effects are typically negligible by spring/summer because usually the ONI itself is pretty close to zero then. Last year was an anomaly, and I'm not prepared to speak on whether the El Nino last spring had anything to do with the very wet spring and above average tornado activity, but having an |ONI| value > 0.5 during the spring and summer is just plain rare to begin with (see how many years have a red number during the AMJ and MJJ boxes in this table: http://www.cpc.ncep.noaa.gov/products/analysis_monitoring/ensostuff/ensoyears.shtml)
3) Usually, for seasonal scale predictability, more signal can be found in weather elements that are highly impacted by ENSO conditions, like precipitation, but again, antecedent precipitation, since it affects things like soil moisture and sensible/latent heat fluxes later on.

 

[Michael Snyder]

2) ENSO effects are typically negligible by spring/summer because usually the ONI itself is pretty close to zero then. Last year was an anomaly, and I'm not prepared to speak on whether the El Nino last spring had anything to do with the very wet spring and above average tornado activity, but having an |ONI| value > 0.5 during the spring and summer is just plain rare to begin with (see how many years have a red number during the AMJ and MJJ boxes in this table: http://www.cpc.ncep.noaa.gov/products/analysis_monitoring/ensostuff/ensoyears.shtml)

Yes-
For example during the 1985-2014 period for Mar-Jul

There were 87 months between 0.4 and -0.4,
35 months between -0.5 and -1.3,
and only 28 months with 0.5 to 1.4 ENSO temps!

And this year by sometime in May we will go below 0.5 (predicted) on the ENSO temps after the huge values we had this winter.

 

[Kyle Brittain]

Came across this recording today:

"Predicting Seasonal Tornado Activity", by James B. Elsner (Florida State University)

https://ams.confex.com/ams/96Annual/videogateway.cgi/id/32117?recordingid=32117

 

[Wesley Luginbyhl]

I threw together this graphic comparison a few weeks back for recent El Nino years and for very strong El Nino years. Well coming out of one like we are in this case. I will add that 2003 can be grouped in with the 95, 07, & 10 group. All I remember about 2003 is that it was the last High Risk for the TX Panhandle (May 15) and that OKC had a rough year.

http://owlsp.blogspot.com/2016/01/el-nino-tornadoes-and-2016-chase-season.html

Copy and Paste from blog below:

First off I will say most correlations with how the storm season will go and what El Nino is doing are very weak. This blog is just to point out some of the more interesting correlations with this El Nino and ones of the past. Also my focus is the Southern Plains, West of I-35 and South of I-70/I-80. This is for the Plains chasers.


One of my favorite correlations, that always catches my eye, is the coming out of El Nino and getting to neutral by mid year. Certain years really stick out there - 1995, 2007, 2010. All were huge years for the Southern Plains, especially the the TX Panhandle. 3 years that any chaser would take in a heartbeat. Currently it looks like our El Nino that is ongoing should follow this path and will be neutral by mid year. If anything will speak well about our upcoming season, that should get you excited.

(all images are clickable and will enlarge)









Annual tornado map for 1995:


Annual tornado map for 2007:


Annual tornado map for 2010:



It is about more than just the number of tornadoes, but also the quality of chaseable tornadoes. That includes if they are photogenic and number of tornadoes days. All of those years were great in those aspects. 1995 may have been all in June, but it was a hell of a June. 2007 was done on May 5, but just about every week between end of Feb till then was very active. 2010 waited till late April, but offered numerous tornado days all over the Plains.


There are some other variables at play though, that should be mentioned. 1995, 2007, and 2010 never reached the strength that the 2015-2016 El Nino has. They were all moderate to weak El Ninos. It can definitely be a factor to consider. Other years that have reached the strong category include the springs of '73, '83, and '98. Below are the annual plots for those years.

Annual tornado map for 1973:


Annual tornado map for 1983:


Annual tornado map for 1998:



Those 3 years are not near as promising, especially 1998 which is down right ugly! It just shows how fickle it can be. There are also other patterns that play into or have even bigger effects than El Nino, but overall should spell good news for the Southern Plains. The drought is gone and should not be returning anytime soon with El Nino around. That alone should make for a decent year, but let's hope for a 95, 07, 10 and not a 73, 83, or 98.