Good evening! I’m bringing you a late night look at an article about Tornado Risk Analysis by P. Grady, Andrew Mercer, Jinmu Choi, and Jared Allen, enjoy!
Summary:
We’re introduced really early on in the article to the idea of “Tornado Alley” and the less widely known “dixie alley”. These two “alleys” are believed to be the most tornado susceptible in the United States and while they don’t plan to define or disprove the idea of tornado alley, they instead seek to identify, “the most tornado-prone location in the United States, regardless of intensity, time of day, seasonality, among others”.
To achieve this, the authors use data from 1950 to 2007 from the National Weather Service and by using the kernel method (as explained in other article reviews of mine) to separate the tornado from the latitude and longitude, beginning and end of the storm. They used kernel density estimation to map patterns of density. Further explanation about the kernel method is stated as, “the Epanechnikov kernel assigns higher density values to locations nearer to the tornado event in question and decreasing values at distances farther from the event in question, eventually reaching zero along the outside edge of the kernel diameter,” in showing the various shading symbolizing density of the storms. A method called bootstrapping is also used to remove bias and anomalies within the data.
They illustrate numerous points about the previous occurrences in various locations around the United States and they show paths of storms with their track. After their looking at their averages from the data, the authors conclude that their study along with many others falls short due to flaws within previous data collected from tornadoes. Additionally, they discover that, “despite the identification of a few Southeastern locations with unexpectedly high tornado-day densities . . . it does not appear as though ‘Dixie Alley’ is its own distinct region.” Furthermore they claim that Dixie Alley is not a separate entity from Tornado Alley when looking at tornado likelihood.
3 Things for Further Study:
- Factors contributing to longer paths of tornadoes in the southeast.
- Extenuating circumstances that caused very few north/south oriented tornado paths. (Fig. 6)
- Previous tornado calculation methods – they express that the increase in tornadoes is likely as a result of modernization of technology for reporting; so what did they do before that made us miss so many?
Modern Meteorology:
This establishes that while the common idea will be to still separate tornado alley from dixie alley, but we can now look to understand that this is just a misunderstanding or lack of knowledge. We see how the density of tornadoes is so strong mainly in Mississippi’s Smith County, the ability to have tornadoes and the probability due to the frequency of the occurrence is so much more common in a wide range of land from the great mid-west plains and reaching toward the Appalachian Mountains and as far north as the Dakotas. By looking at tornado density and frequency in an objective manner rather than including useless aspects of data that skew the reality of the frequency.
