There is an obvious increase in tornado frequency between 1950-1999. This could be due to increased detection. Also this could be due to changing climatic conditions. Looking at the raw data we have seen that there are generally less tornadoes in El Nino years compared to La Nina Years. But, since we were unable to get climate data, we were unable to see if the change in the frequency was due to climate factors.
Our data has failed to show a strong correlation in increase in tornado frequency and magnitude during El Nino and La Nina events. P-values indicate that there is some relationship between the frequency of tornadoes in La Nina years between states and the USA. For example, the p-value for rank orders between Texas and Nebraska is .0422therefore there is no significant difference between them.
For El-Nino years, only two of six p-values are under .10, revealing that there is a majority of rank orders which are significantly different. For example, Texas and Nebraska had a p-value of .4902, meaning that there is a significant difference.
In non-El-Nino years, we again saw 2 of 6 tests producing p-values under .10. Therefore, 4 of 6 rank orders produced a significant difference. Texas and Nebraska rank orders had a p-value of .1096, meaning that there is a significant difference in that test.
In non-La-Nina years p-values indicate a strong state and national correlation. Because all but one p-values were under .10, we could reject the null hypothesis that there is a significant difference in rank orders. For example, Texas and Nebraska yielded a p-value of .0719.
For total magnitude counts, we see p-values of less than .10: there is no significant difference in rank orders. Texas and Nebraska yielded a .0253 p-value.
For magnitude in El-Nino years p-values were also less than .10 and we see no significant difference in rank orders. Texas and Nebraska yielded a .0350 p-value.
For magnitude in La-Nina years p-values were again less than .10 and there is no significant difference in rank orders. Texas and Nebraska yielded a .0639 p-value.
Suggestions for future research would include an investigation into tornado frequency by latitude, rather than region alone, in El Nino and La Nina years. Because of the latitudinal shift in the jet stream, an important ingredient in tornado formation, we believe this could reveal some interesting results.
See our first presentation when we present tornadoes to the class.
See our tornado powerpoint presentatin where we present our final results.