An Increase in Significant Tornado Frequency
in the Paducah Area


"Tornado Alley" is defined by the American Meteorological Society as: "A term often used by the media to denote a zone in the Great Plains region of the central United States, often a north-south oriented region centered on north Texas, Oklahoma, Kansas, and Nebraska, where tornadoes are most frequent. Since statistics are variable on all timescales, the term has little scientific value."(1) However, over the past decade, a noticeable increase in tornado activity has been seen outside of this area, spreading farther east into the Mid-Mississippi Valley and Mid-South. This increase in tornado activity seems to center around and south of our station in Paducah, KY, which covers parts of southern IL, southeast MO, west KY, west TN, and a small part of southwest IN. To put this increase in perspective, over the past decade, a region from western Kentucky and western and middle Tennessee has experienced more EF-4 tornadoes than any other location in the country. Also, since the Paducah NWS began keeping records in 1950, over 1/3 of the recorded EF-2 and above tornadoes have occurred within the past 15 years. (6) This data bears asking the questions, "Is there a new Tornado Alley?" In this paper I will attempt to answer that question, as well as explain some possible reasons for this perceived shift in tornadic activity, and discuss how it affects the work of local forecasters and media personnel.

Over the past decade, 16 fatal tornado events have occurred across our area, with 85 total fatalities. All but two of these events have been strong or violent EF-3 or EF-4 tornadoes, and nearly of these events occurred after dark. Also, half of these fatal events occurred outside of the "typical" tornado season of March-June. Some of the notable events within this period include:
--Jan. 17, 1999-Jackson/Madison Co., TN; EF-4; 6 dead
--May 4, 2003-Jackson/Madison Co., TN; EF-4; 11 dead
--May 6, 2003-Grand Chain/Massac Co., IL; EF-4; 2 dead
--Nov. 6, 2005-Evansville, IN; EF-3; 25 dead
--April 2, 2006-Caruthersville, MO, & Dyer/Gibson Co., TN; EF-3; 26 dead
--Feb. 5, 2008-Super Tuesday Outbreak: Greenville, KY; EF-3; 3 dead & Jackson, TN; EF-4; 51 injured at Union University

A recent study by Boruff,, published in 2003, used a 50 year record to determine trends in tornado hazards, which were defined as any reported tornado that caused injuries, deaths, or any amount of economic loss. The study showed both a steady decline in injuries and fatalities overall during the period, but noted a significant increase in the number of tornado hazards in past decades. Although this may seem alarming, one may quickly deduce that this increase can be attributed to vast improvements in observation, reporting, and documentation. In addition, it was found that the frequency of EF-2 or higher tornadoes has remained relatively steady, and that the increases in tornado-related hazards were not due to weather changes, but instead were attributed to human behaviors and locations. In examining the overall trends in tornado hazards geographically, however, a noticeable trend did occur. The overall number of tornado hazards for each decade was calculated and a "centroid" location was assigned, and that location slipped farther to the southeast every decade. Their findings suggested that the most recent tornado hazard centroid would be located near Poplar Bluff, MO, which would support the idea that the density of tornado hazards has expanded eastward from the traditional "Tornado Alley". (2)

In addition to the geographical distribution of tornadoes, one must also a look at trends in tornado fatalities. A study by Walker Ashley of Northern Illinois University shows troublesome findings pertaining to tornado fatalities in the southern US. Ashley used the NCDC's Storm Data database and a long-term study of US tornadoes by Thomas Grazulis to examine fatalities from 1880-2005. A 60km x 60km grid of tornadoes and tornado fatalities from 1950-2005 reveals that even though the number of tornadoes are greater through the traditional Tornado Alley area, the majority of fatal tornado events occurred mainly outside of that region in the lower-Arkansas, Tennessee, and lower-Mississippi River valleys. Based on the long-term data, it doesn't seem likely that this distribution could be attributed to a climatological shift in tornado occurrences, but instead a combination of climate and non-climate related factors can help explain this distribution. These factors include: 1) Data shows that this area sees a greater proportion of F2 and above tornadoes within the overall number of tornado reports. 2) A larger number of tornadoes in this region occur during the cool and transitional seasons when days are shorter, increasing the likelihood of nighttime tornadoes. Also, this means more storms occur during seasons when the overall atmospheric flow is higher, which leads to faster-moving storms and less lead-time with warnings.

3) A larger percentage of mobile homes and lower-stock housing materials are used in this region that any other in the nation. From 1985-2005 mobile home fatalities accounted for 44% of all fatalities, and in the southeast, the percentage of mobile homes averages over 20% in most counties. 4) This region of the country is heavily forested, which greatly reduces the visibility of approaching storms. 5) Low-level humidities are much higher in this region, which leads to lower-based storms and higher percentages of rain-wrapped tornadoes, enhancing the difficulties in spotting tornadoes. 6) The overall population density in the southeast is greater than in the traditional Tornado Alley. 7) A clear tornado-season is not well defined, with a slightly lower, but more consistent risk of tornadic thunderstorms throughout the year. This may lead to instances of people being caught off-guard in outbreaks outside of peak months. 8) Several demographic trends, including a higher percentage of elderly and impoverished persons may enhance this region's vulnerability. (2) Another question that must be taken into consideration is whether these trends are consistent through time, or if there may be climatological influences that lead to times of a higher or lower threat of tornadoes in this region. One suggestion that has been examined is the influence of El-Nino-Southern Oscillation (ENSO) on the weather patterns of the US, especially during the cool seasons.

A study by Cook and Schaefer suggests that ENSO plays a large role in the seasonal positioning of the jet stream, which is a major player in where conditions favorable for thunderstorm development will set up. As a whole, results from this study show that the number of cool-season tornadoes is unchanged regardless of ENSO phase, however, the geographical distribution of these tornadoes do fluctuate with the ENSO phase. The findings show that during La Nina years, organized tornado activity lies within a band stretching from Louisiana northeast through west Tennessee and Kentucky, southern Illinois and Indiana, into Michigan. During neutral-phase years, a band of activity is found in a west to east zone from Oklahoma and Kansas through Tennessee and Georgia, and during El Nino years, most of the activity is suppressed to an area along the Gulf Coast. Also, the study found that the frequency of strong and long-track tornadoes were higher during La Nina and neutral-phase years, which may help to explain the increased number of fatal tornadoes in the Mid-South. It is important to note, however, that ENSO phase can only be viewed as a background influence, and not a significant control of tornado outbreaks. Large-scale synoptic features, such as jet-stream positioning, also has to be accompanied by numerous other mesoscale factors, such as the proper location of contrasting temperatures, low-level moisture, favorable wind fields, helicity, instability, and even smaller-mesoscale influences that allow for localized updraft development. Nonetheless, seasonal outlooks that reveal the upcoming ENSO phase should be looked at to gain an idea of overall trends possible over a season. It may be worth noting that the 1925 Tri-State Tornado, the 1965 Palm Sunday Outbreak, and the 1974 Super Outbreak all occurred during La Nina phases. (4)

As a forecaster and member of a media outlet, an examination of this data (as well as the experience of seeing EF-4 tornado damage first-hand) reinforces the importance of providing an accurate assessment of possible severe weather situations and the need for timely warnings to the public, especially given the lack of visibility of storms and the human demographics discussed earlier. Also, due to the enhanced frequency of deadly nighttime tornadoes, our station has launched and maintained a campaign to provide viewers with the opportunity to purchase NOAA weather radios for a discounted price. This campaign began after the Nov. 6, 2005, Evansville tornado, which struck a mobile home park at around 2 am, killing 25. Since then, we have sold nearly 20,000 radios, which will provide warning to people that may never know of approaching storms otherwise. Many other campaigns throughout our area are conducted by the National Weather Service in order to raise awareness of the year-round danger that tornadoes can pose.

As important as it is to raise the awareness of the public to the threat of tornadoes, a problem associated with that can be finding the balance between waiting to move on a warning for a storm and over-warning. It seems that over the past decade, with the increase in tornado reports and fatalities, has also come a tendency to "blanket warn" on storms, causing large numbers of people to be under warning, with the great majority of them failing to see even severe thunderstorm conditions. One helpful solution has been the shift from county-based warnings, to the storm-based warning polygons. It will take some time for viewers to get used to this method of warning, but overall will help to reduce the total number of "false alarms" and will eventually decrease the public's tendency to view a warning as a forecaster "crying-wolf".

In summary, an examination of multiple studies shows that an overall shift in the number of tornadoes outside of the traditional Tornado Alley does not appear to be occurring, however, the number of tornado-related fatalities has been shown to occur in a region including parts of Arkansas, Tennessee, Mississippi, and Alabama. This, combined with improvements in the detection and documentation of tornadoes, may give the perception that a shift in tornado frequency has occurred. Some media outlets and forecasters in this region have even gone so far as to proclaim the Mid-South as the "New Tornado Alley" or "Killer Tornado Alley" (6). One meteorologist, Alan Moller of the Fort Worth NWS office, suggests that the number of tornado-related fatalities outside of the traditional Tornado Alley may warrant either the expansion of Tornado Alley or the discontinuation of the term. It has been noted that tornado preparedness and awareness is higher within the traditional Tornado Alley, therefore the thought would be that an expansion of that area may also lead to heightened awareness, which may in turn lead to a lower death toll within the debated area (5).


(1) American Meteorological Society, Glossary of Meteorology.
(2) Ashley, W. 2007: Spatial and Temporal Analysis of Tornado Fatalities in the United States:
1880-2005. Weather and Forecasting, 22, 1214-1228.
(3) Boruff, B. J.; Easoz, J. A.; Jones, S. D.; Landry, H. R.; Mitchem, J. D.; Cutter, S. L. 2003:
Tornado Hazards in the United States. Climate Research, 24, 103-117.
(4) Cook, A. R. & Schaefer, J. T. 2007: The Relation of El Nino Southern Oscillation (ENSO) to
Winter Tornado Outbreaks. NOAA/NWS Storm Prediction Center.
(5) Okada, B. 2007: Is It Time to Widen Tornado Alley? McClatchy-Tribune.
(6) WZTV, Fox 17 Nashville, Report by Scott Couch.