Dry autumns and winters may lead to fewer tornadoes

[Jason Foster]

Not exactly the most reliable source, but this was posted on another forum, and thought I'd share, see if anyone is familiar with the study, or can discuss it further.

(FULL STORY LINK) from June 2009

Dry autumns and winters may lead to fewer tornadoes in the spring, according to new analysis of long-term data

Athens, Ga. – Global warming will likely mean more unpredictable weather, scientists say, and a new study by researchers at the University of Georgia pins down, possibly for the first time, how drought conditions in an area’s fall and winter may effect tornado activity the following spring.

The study, published today in the journal Environmental Research Letters, is specific to Georgia and the Southeast, but further study could reveal patterns that might make this more general—including the already tornado-prone Great Plains.

.....See link above for full story

Writer: Philip Lee Williams, 706/542-8501, [email protected]

Contact: Marshall Shepherd, 706/542-0517, [email protected]

Source: University of Georgia

 

[Patrick Marsh]

Just how the connection works between fall-winter rainfall and spring tornado seasons remains unclear. One possibility is that the atmosphere uses soil moisture “memoryâ€￾ from the fall and winter to modify conditions suitable for severe weather. A related hypothesis is related to “soil moistureâ€￾ pockets and storm initiation.

Um..how about this connection...

During "wet" fall-winters (which tend to occur in El Nino years), the upper-level storm track is most favorable for storm systems to move through the southeast (also tends to occur in El Nino years). This pattern tends to persist into the spring. Increased storm system days will lead to an increase in potential tornado days. Assuming that the ratio from potential tornado days to tornado days is relatively constant, then more storm system days would equal more tornado days.

I hope the actual paper addresses this issue as well as several others I have from the write-up...

 

[Chris Nuttall]

I can see how this could be applied on a regional scale. For example, the last two fall and winters have been pretty dry in the TX and OK Panhandles, leading to very dry soils. The lack of evapotranspiration has allowed to dryline to develop much farther east on most days, and the Panhandles have had below normal numbers for tornadoes in 2008 and 2009. Of course, there is also probably an El Nino and La Nina connection to the wet/dry winters, like what Patrick mentioned.

However, total tornado numbers for any given year are one of those stats, like rainfall, that can easily be skewed by one or two major events. Here's a link to the full article:
http://www.iop.org/EJ/article/-search=66527503.1/1748-9326/4/2/024012/erl9_2_024012.html

 

[cdcollura]

Good day all,

The BIG word here is "un-predictable".

For example, May is supposed to be our big month chasing in the central US for tornadoes ... But during these "un-predictable" times - Storms affect areas OUTSIDE tornado alley (usually with a stout ridge there in the middle in May).

The tropics, for example ... Grace forms off Europe, yet the Caribbean / Gulf of Mexico remains dead as a doornail - Again, storms where you NORMALLY do not get them.

I think the good ol'e pattern with a busy May and tropical season closer to Florida as "usual" is desired on my behalf.

I hear many people at my job, and family, telling me "crazy weather this year, huh?" ... Yet this is ONE of my worst chase years. Kinda contradicting.

The dry fall and winter theory could be correct, but this year we are in a mild El Nino pattern, which would be a wetter winter. El Nino is responsible for ridging the US plains in lat May 2009, as well as quelling the atlantic tropical season due to the strong flow aloft (shear) ... These are, as frustrating as they are, signature effects of the El Nino cycle.

 

[Patrick Marsh]

I've read the paper now. (Thanks, Chris, for a link to the full article.) A couple of points:

1. Correlation does not equal causation.

2. Nothing was mentioned about the number of "storm" days or even "precipitation" days. I'd argue that while in a drought (such as 2006-2008), then of course you are going to have fewer tornadoes- you don't have as many precipitation events that could possibly yield tornadoes! I'd need to see something said about the relationship between the number of precipitation days to number of tornado days before considering their premise that fall and winter drought conditions limit the number of tornadoes in the spring. I'm tempted to read this paper more along the lines of saying, "When fall and winter are drought-like, then the spring is more likely to be drought-like, which results in lower tornado numbers."

The old adage, "When in drought, leave it out," certainly could apply here...

 

[Chris Nuttall]

I have had a little bit of correspondence with Dr. Shepherd on this study. With his permission, here is a response to some of the discussion on this thread that he wished to have posted. Thanks goes to Patrick to relaying his email to me.

Thanks for discussing our recent paper on this topic, we were quite surprised when we found this preliminary finding and are quite careful to note that this paper does not investigate mechanisms. We speculate on a soil moisture memory but I also currently have a graduate student conducting a more robust analysis looking at synoptic pattern response as well. Bottom line, there is something that may be lurking in Fall-Winter rainfall patterns that lend some seasonal predictability, but it is way too early to say with certainty. The other key takeaway from our paper is that we only saw suppressed tornadic activity under drought conditions. We did not necessarily see more tornadoes under extremely wet conditions. The study should be extended and conducted under an array of conditions and in more tornado prone regions, which we are currently undertaking. We have talked with Harold Brooks and Joe Schaeffer (Joe and I were on AMS Council together) about collaborating on this work and they were quite interested

Note, we did not try to extend these results are read too much into them at this point, if you read the paper you will see that we are very cautious in interpreting this very limited study on tornado days. Of course, correlation is not causation and we state that in our discussion as well.

Great discussions..

Marshall Shepherd