Paper No. 291-12
Presentation Time: 9:00 AM-6:30 PM
ANALYSIS OF F4 AND F5 TORNADO DISTRIBUTION IN THE UNITED STATES FROM 1900-2014
Violent (F4 and F5) tornadoes in the United States are most prevalent in portions of the Midwest and South known respectively as “Tornado Alley” and “Dixie Alley.” In an attempt to determine whether any long-term trends exist in where violent tornadoes occur, we have conducted spatial analysis of all 1,055 violent tornadoes in the United States from 1900 to 2014. The tracks of tornadoes were mapped using information from Grazulis (1993) and from National Weather Service damage surveys. The tracks were converted into polylines and divided by county using ArcGIS. The counts of county segments were used to generate heatmaps from county centroids using the Kriging method. Grazulis found that through much of the twentieth century a fairly dense distribution of farms in the United States has served as a relatively reliable indicator of tornado distribution. Breaking data from the full 115-year period into five equal-length periods of 23 years reveals numerous randomly distributed hot spots (areas of high tornadic activity). Most of the tornadoes associated with a given hot spot will have occurred on the same date, indicating that these hot spots represent individual tornado outbreaks. Most of the variation between periods therefore appears to be random statistical noise. One possible exception is the period of 1969-1991, which shows an apparent eastward shift in the occurrence of violent tornadoes, due largely to several major outbreaks that occurred during that period. The greatest increase in activity occurred in an area extending from the Mississippi River to the Carolinas and Pennsylvania. In other periods the greatest activity largely remained west of the Mississippi. This shift occurred with a reduction in the number of violent tornadoes in the United States that occurred in the mid to late 1970s and continued into the twenty-first century. The change in tornadic activity temporally coincides with a shift in the long-term pattern of the El Niño Southern Oscillation (ENSO) with El Niño events becoming stronger and more frequent and La Niña events generally becoming weaker and less frequent. This apparent correlation suggests the possibility of a long-term climatic trend. Further analysis will divide tornado paths using a uniform grid instead of counties to eliminate any effects of the non-uniform sizes and shapes of counties.