Yes, the title isn’t exactly gasp-worthy. Nor is it a ground-breaking new idea. I know. But What I found out was interesting!
For those of you who frequent this site – and there aren’t many of you – you know that I’ve been rather preoccupied the last week or so digging through data from the National Weather Service office in Amarillo, Texas.
I was looking at data from 1997 through 2013. I examined global sea surface temperatures, monthly rainfall for Amarillo and spring tornado reports.
I was trying to keep my side project “on the DL” as the kids say. No real reason, I just didn’t want to “jinx” myself, I guess.
Yesterday, I completed the statistical analysis of all of the data and I’m preparing a quick write up for how the numbers all add up. But before I start writing, I wanted to give you guys a look at the data I was sifting through (above), as well as, a few things I learned.
Things I learned:
- After looking at everything, there is a potential relationship between the Gulf of Mexico SST, Gulf of Alaska SST, ENSO, October to April rainfall and the number of tornadoes across the Texas and Oklahoma panhandles
- Strong La Nina and El Nino show “lag time” in changing the number of tornadoes
- Assigning values derived from previous researchers to each variable, a given year’s tornado activity (“above average” or “below average”) can be accurately predicted 76-percent of the time
I want to caution everyone that this little project is not intended to be “scholarly research” at all. This is just something that sparked my curiosity one day and I wanted to investigate. It should probably not be cited or used in future research – of any kind.
The main goal was to see if there was a relationship between the variables, and if so, to what extent. Beyond that, perhaps a forecaster could keep it in the back of his or her mind when looking at a spring outlook – but nothing more.
Now, it’s time to start typing…