Our theories about Nino and Nina and tornadoes

[Andrea Griffa]

Our theories about Nino-Nina and tornadoes

Hi all
I'm starting a new thread to discuss our theories about Nino-Nina and the connection with tornadic activity in the Plains. So I exhort all of you who have got some thoughts about that to write down.

As regards me I've made a little study in order to find a connection between Nina-Nino and tornadoes: I considered this 2 links: first, Tornado Statistics by Year and Month 1950 - 1997.


http://www.disastercenter.com/tornado/tornad2.htm

Then, the ONI index map:


http://www.cpc.ncep.noaa.gov/products/analysis_monitoring/ensostuff/ensoyears.shtml

I've never heard about conclusive studies that prove a clear correlation between Nino-Nina and Tornadoes: anyway I've got an idea; you could tell me this is pseudoscientific anyway I try to find something that could be useful for the forecasting of the chasing seasons.


As to me we could individuate some patterns: making a little study about ONI index(Changes to the Oceanic Niño Index) I found that the best pattern to have an high number of tornadoes during the month of April and May is a weak-moderate Nino during the firs 6 months of the year and a moderate-strong Nino in the next 6 months. The typical example is 2004, 1982,1991, 1994. This idea is supported by the fact that in those years the number of tornadoes has been high and during the months of April and May we had on average more tornadoes than other years with a different ENSO pattern.


Now talking about Nina, watching at the last 40 years we had 1974-1975-1985-1989-1999-2000.

74 and 75 had strong Nina : 74 was the year of the Superoutbreak and both the years had a moderate high number of tornadoes during April and May (267 and 144;108 and 188).

1985 had a good number of tornadoes during April and May (134 and 182).

1989 was a very good year for May (82 and 231).

1999 was a very good year for May even if I didn't find any other documentation..If you find the numers, post'em.

2000 was a moderate year with about 170 tornadoes in May and 75 in April but without relevant episodes.

Finally it seems there's a correlation between strong Nina and rich in tornadoes episodes during the months of April and May.

As regards this year we have moderate-strong La Nina and at this point I hope it will remain in this way till spring and let's see what will it be: if the theory is true I should expect an outbreak after the first half of April and maybe another one within may 5, and a moderate/high numer of tornadoes in April-May.

-Also I found that 1995,1998 and 2003, (probably those are the most tornadoes producer years,in April and May) shows a similar ONI pattern with neutral April-May (ONI around 0) preceded by 11-13 months with strong Nina. Incredible how Nature repeats itself.

-Moreover another best tornado producer pattern is 2004 like pattern: first 4-5 months with weak Nino followed by 11-13 months with moderate-strong Nino(2004, 1982,1991, 1994).

Now let me know what do you think about this theory and tell me if you have some other theories.

 

[Chad Cowan]

Good topic, Andrea. I've been researching this a bit lately and my early thinking is that ENSO has more of an impact on the geographical concentration of tornadoes than the frequency. This is due to the impact of SSTs on the positioning of the subtropical and polar jet streams. A typical La Nina pattern will generally weaken the subtropical jet and place it further north than 'normal', while a typical El Nino pattern will generally place the polar jet further south.

Sources/ reading material:
http://ams.confex.com/ams/pdfpapers/115322.pdf
http://www.spc.noaa.gov/publications/schaefer/pacsst.htm

 

[Niccolò Ubalducci]

We are waiting for answer!

 

[Paul Knightley]

I have read forecasts which predict La Nina to weaken considerably through the next couple of months.

 

[Alexandre Aguiar]

Great reading:

http://climatesci.org/2008/02/12/relationship-of-family-tornado-outbreaks-to-la-nina/

 

[Verne Carlson]

Great article Alexandre, I hope it's correct - the data seems sound although I'd like to see an update using data from the 90's. The paper was published in 1993.

I have read forecasts which predict La Nina to weaken considerably through the next couple of months.

Paul, if I read the paper correctly the effects on the continental US lag 3-5 months behind the actual La Nina event so we would still be under it's effects for the 2008 season.

Also note in the papers data that of the La Nina events you have this monthly breakdown (excluding fall and winter)

# of Outbreaks
----------
Mar - 1
Apr - 4
May - 2
June - 1

April is the big month here for tornado outbreaks during La Nina years!

 

[Shane Young]

I had a long discussion today with some friends about La Nina pattern's correlation to high tornado counts / geographical distribution / significant tornadoes and it led me here...


We were mainly looking at
http://www.cpc.ncep.noaa.gov/products/analysis_monitoring/ensostuff/ensoyears.shtml and StormData. I hadn't seen the monthly/yearly page you gave Andrea, that would'a been helpful!
Also may want to use http://www.ncdc.noaa.gov/img/climate/research/tornado/tornadotrend.jpg for looking at when violent tornadoes occurred. I really didn't.
If you use 3/4 index of <= -0.5 in spring, you get 1971, 1973-1976, 1989, 1999, and 2000... with 1988 being all but officially a La Nina.
If you use 3/4 index of <= -0.5 in winter, you add 1996 and 72 (and remove 88).


There were 1169 tornadoes in 2000.

It just looks like there's no real signal here. 92 had 1297 tornadoes... it was a strong El Nino year. It's not like El Nino years have much of a drop off...
And time-wise, same deal. 1991 had one of the busiest March-May periods (Andover anyone?!?)... and it was shifting from weak El Nino to strong.
And haven't I heard that 1988 is infamous for it's SDS?


I didn't look too much into location by year... except to note that while 99 had the record outbreak, things were relatively quite quiet nationwide in the May 3 - May 10 period of May in 85, 89, and 2000 (30, 43, 30 tornadoes) and were mainly maximized in the northern and eastern US. Meanwhile 2003 was a diminishing El Nino, and we probably all remember that! The northern Plains pretty consistently had 2-3 small tornado outbreaks in the May 25 - June 15 period during those 4 LN years.


There may be weak signals out there. I do note that there are only four times in the past 30 years where a La Nina was seen ramping up quickly during the spring/early summer..... those specific springs seemed like maybe they were just a tad quiet... but the next years were 1965, 1974, 1989 (nothing), and 1999.
Maybe a signal. Maybe just an artifact of a small sample size. Either way, it's mute for this year because the La Nina developed more towards fall last year.
The direction ENSO is going... either in December... or in March-May... may have an impact, I didn't check.
Then again, the 2nd or 3rd derivative may show the strongest signal for all I know. Somehow I doubt they are strongly significant.


I do note that the major forecast that was put in papers/articles after 1999's La Nino was that the deep south would have many more tornadoes in the winter. That definitely happened this year!
Maybe there is a much greater impact when the cold air mass is more deeper/consistent (in the winter)? That does interest me a bit more.



The point here, to me, is the same deal I made back before last summer about making seasonal hurricane forecasts.
It seems pretty much useless to try to forecast whether this will be an active year based upon climatological features like ENSO. There may be a slightly above or below average percentage chance for something based on records... much like we see in CPC precipitation forecasts... but I don't think it [ENSO] should really be a part of how people do things this spring. I'm planning a late-spring/early-summer chase, and I'm really not concerned with ENSO!

It's fun guessing. And I tend to trust those who've been around and developed the real "feel" for the season (another thread on ENSO I came across was http://www.stormtrack.org/forum/showthread.php?t=14847) like Gene Moore... so I don't discount the ENSO effect too casually... but still, it's so mesoscale/short-term-system based... I really don't see any use in looking at ENSO.



I welcome to see some major statistics that disagree! It'd make planning things a lot easier! But I think we're off looking for words in a 3 year old's scribbling at the moment with all this!

 

[Andrea Griffa]

I had a long discussion today with some friends about La Nina pattern's correlation to high tornado counts / geographical distribution / significant tornadoes and it led me here...

Hi Shane
Good considerations.If you're interested to find a possible ONI pattern and you're not too much skeptical I suggest you to see the monthly satistics with ONI index and try to find a possible correlation. Try to esaminate the consideration I expressed at the first part of the thread, and if you pay attention you can find a kernel of thruth about my ONI patterns.

Thanks a lot about the link of Gene's discussion about Nina; I report the Gene's considerations that are very important as well. I encourage Gene to come in this thread and speak about this arguement.

Gene's point is the one I suspected overall when he speaks about tornadoes more common near and east of the Misissipi Valley and eastern USA and about winter outbreaks. The key of La Nina year is to catch the right outbreak and give the best that you can in those chasing days.

My thoughts on a typical La Nina spring include some research and my personal experience. For me La Nina is the least productive for tornadoes, although I've had some very spectacular chase days in La Nina years.

- More drylines - hot and dry SW US - stronger cap firing late or east
- Less chase days and opportunities on the classic plains and high plains
- Tornadoes more common near and east of the Mississippi and eastern US
- More night tornadoes over plains (some research to support this but not conclusive) result of stronger cap
- Cold season tornado events and outbreaks (already seen this recently)
- Big Outbreak - some of the biggest outbreaks have occurred during La Nina springs. If this big day occurs and you miss it you may miss much of the season. Examples: 1974 April 3/4 Midwest US - 1974 June 8th Oklahoma. 1999 May 3rd.

Generally my La Nina years are the leanest for tornadoes unless I nail the outbreak. Outside of those years my worst chase years have been La Nina springs such as 1988, 2000.

Some helpful links for La Nina:
http://ggweather.com/enso/years.htm
Specifically 1998:
http://www.pmel.noaa.gov/tao/elnino/...a-pacific.html
These maps I think show typical La Nina better::
http://www.cpc.ncep.noaa.gov/product.../mamninapr.gif
http://www.cpc.ncep.noaa.gov/product.../mamninatr.gif
From this page:
http://www.cpc.ncep.noaa.gov/product...ina/index.html
Another link in plain language:
http://www.stormfax.com/elnino.htm

Also see Stormtrack for La Nina discussions... there was one in 06.

 

[Chad Cowan]

Here are some composites I made of the SOI and tornado totals by year. Both tornado data are the same, just presented differently. A positive SOI (blue) indicates the presence of La Nina while a negative SOI (red) indicates the presence of El Nino.

Sources:
http://www.hprcc.unl.edu/nebraska/U_S_SEVERE.html
http://www.john-daly.com/elnino.htm

 

[Andrea Griffa]

Retrieving the arguement, the 2008 average of dynamical models forecast about SST anomalies is this:


-1.4 -1.2 -1 -0.8 -0.7 -0.6 -0.6 -0.5 -0.4

I found that the most similar La Nina year to this pattern is 1985(and 1974).

0.8 -0.8 -0.8 -0.6 -0.6 -0.5 -0.6 -0.4 -0.4

I found some stories about 1985:

http://www.stormtrack.org/archive/x09.htm

Another year that could be very very similar is 74'


-1.2 -1.1 -0.9 -0.7 -0.5 -0.4 -0.5 -0.7 -0.8

Or 2000

-1.0 -0.8 -0.6 -0.6 -0.4 -0.4 -0.4 -0.5 -0.7

 

[Scott Lincoln]

I actually just finished up this mini GIS project on my own regarding possible La Nina effects on intensity of tornadoes. It still needs some work to be considered totally scientific, as I've had some issues getting some of the different grids to line up properly. I did find it interesting none-the-less that there is almost a region-wide increase in the intensity of tornadoes in La Nina months compared to all months.
I'll have to look into some of these other papers/sources.

 

[Gene Moore]

I actually just finished up this mini GIS project on my own regarding possible La Nina effects on intensity of tornadoes. It still needs some work to be considered totally scientific, as I've had some issues getting some of the different grids to line up properly. I did find it interesting none-the-less that there is almost a region-wide increase in the intensity of tornadoes in La Nina months compared to all months. I'll have to look into some of these other papers/sources.

Since I've chased years where there were almost no tornadoes in the areas you show a big increase I'm anxious to see the rest of this research. I wonder if it's skewed by the big years, like 1999 as opposed to 1988? Looks like those wanting more tornadoes will get their wish, but so far this LaNina appears to be following the old rules....night events, more east of the Mississippi etc.

 

[Shane Young]

Some interesting stuff!

I guess I think about everything too much from the perspective of a chaser sometimes, especially when it comes to chat on StormTrack (for instance, I was horribly upset that a moderate risk was posted on Tuesday for the Carolinas. There was very little chance of tornadoes, especially of the isolated, supercellular, chaseable variety. Of course the moderate did verify for the winds, which was really the focal point of the MDT (though it did have an excessive 15% tor), but I spent the whole afternoon decrying the risk. The point being: all I seem to think about sometimes is chasing!).

Anyways, it gets so muddled when you get into statistics. Just about anything can be found if you try hard enough (the inverse correlation between pirate ships and global temperature!). From Chad's images, it actually looks to me to suggest El Nino years typically have more tornadoes!?! Then you've got Scott's work showing an increase in intensity in La Nina years. And I've seen the data showing more tornadoes for La Nina too.

I guess my point/thinking still is from the chasing perspective; that all this therefore is just not important. There will be setups and tornadoes this year, like all others.

I tend to be so very one-track minded! Because the weak signals don't have any significant effect on chasing, I sometimes end up not paying any attention at all. But it's an interesting topic nonetheless. Hopefully I'll get some more time in the future and can put do some rigorous statistical analysis of my own and see what comes out... specifically in terms of region, time, and month.

But in the mean time, I continue to enjoy reading what everyone else is finding!

 

[Aaron Kennedy]

It still needs some work to be considered totally scientific, as I've had some issues getting some of the different grids to line up properly.

What's your methodology? I'll save any comments until I see that =) There are a variety of papers that look at El Nino/La Nina's influences on tornadoes. It seems like almost all of them show crazy variations from CWA to CWA or county to county, etc. This raises huge warning flags to me. Teleconnections should cause a slight change in probaility over regions... not specific counties. As Gene pointed out, many of these studies are flawed in the number of cases considered... ONE outbreak can significantly bias results.

 

[Andrea Griffa]

Hi all



Now talking about Nina, watching at the last 40 years we had 1974-1975-1985-1989-1999-2000.

74 and 75 had strong Nina : 74 was the year of the Superoutbreak and both the years had a moderate high number of tornadoes during April and May (267 and 144;108 and 188).

1985 had a good number of tornadoes during April and May (134 and 182).

1989 was a very good year for May (82 and 231).

1999 was a very good year for May even if I didn't find any other documentation..If you find the numers, post'em.

2000 was a moderate year with about 170 tornadoes in May and 75 in April but without relevant episodes.

Finally it seems there's a correlation between strong Nina and rich in tornadoes episodes during the months of April and May.

As regards this year we have moderate-strong La Nina and at this point I hope it will remain in this way till spring and let's see what will it be: if the theory is true I should expect an outbreak after the first half of April and maybe another one within may 5, and a moderate/high numer of tornadoes in April-May.

Ok, now it's time to see if we find some weather proofs that this pattern could be really similar to 1974' one: synoptic conditions show the next days with 3 moderate risks on stream and one of those (April 10) presents very similar features that could remind 1974 pattern (difference of 200km for convective initiation, so far) that is what I was expecting for this period. Oni index forecast shows a similarity 1974' Oni pattern (as I found in the last threads).


-1.2 -1.1 -0.9 -0.7 -0.5 -0.4 -0.5 -0.7 -0.8

At this point I don't expect necessary a superoutbreak like 74 (that is impossible to repeat in my opinion) but at least a number of tornado > 200 during April, the major part collocated not in Tornado Alley but in the central East States(just like Nina teaches).
If this things happen we should start to take at least in a little consideration that it could be true that Oni Index Correlation could predict(at least in a modest percentage) the possible behaviour of a chasing season and this could be the start of other stormtrack research and studies(scientists, don't kill me)

Now, let see if the story repeats itself. Good luck to all chasers that will be in the right place and I pray to have less injuries as possible in the populations.