1. Introduction
Debates have been ongoing about the location of the most tornado-prone regions of the United States over the past 10 yr or so. For many years, “tornado alley,” a region in the Great Plains extending from northern Texas to Iowa, was considered to be the area with the greatest tornado risk in the country. However, the idea of another local maximum of tornadoes in the southeastern United States was first mentioned by Alan Pearson in 1971 (Gagan et al. 2010), and further shown to some degree by Concannon et al. (2000), Ashley (2007), Dixon et al. (2011), Smith et al. (2012), and Kellogg and Forbes (2013).
Researchers have also used various ways of defining tornado-prone areas and, by implication, tornado risk. Using data from what eventually became the Storm Prediction Center (SPC) tornado database (e.g., Schaefer and Edwards 1999), Kelly et al. (1978) reported that the area of the United States with the highest tornado frequency was the Great Plains, or the classic tornado alley [as mapped by Schaefer et al. (1980) and from McNulty (1995); Fig. 1a]. Brooks et al. (2003), using tornado days and limiting the risk areas to those with tornado seasons that are consistent by time of year, suggested that the primary tornado alley in the United States extends from Texas to North Dakota, similar to and slightly northwest of the classic tornado alley (Fig. 1b). Concannon et al. (2000) examined days with significant tornadoes [those rated on the Fujita scale as category 2 (F2) or higher] from 1921 through 1995, and found an L shaped region of maximum tornado risk that included most of the classic tornado alley from Schaefer et al. (1980), then extended eastward through Arkansas, Mississippi, and Alabama (Fig. 1c). Smith et al. (2012), using a more limited dataset (2003–11) including tornadoes of all intensities, suggest that the southern part of the tornado-prone area shown by Concannon et al. (2000), from Oklahoma to Alabama, has the highest tornado risk, with a maximum over parts of Mississippi and Alabama. However, this small dataset was probably overly influenced by the 27 April 2011 outbreak (Fig. 1d). Ashley (2007), using only significant tornadoes during the years 1950–2004, and Carbin et al. (2012), examining the return frequency of significant tornadoes from 1961 through 2010, also showed that the region from Oklahoma to Alabama has the highest tornado risk (Figs. 1e and 1f).
(a) Annual tornado frequency for all tornadoes from Schaefer et al. (1980), shown in McNulty (1995), based on records before 1980. (b) Tornado occurrence based on frequency of all tornadoes, 1980–99, combined with a measure of seasonality of tornadoes (Brooks et al. 2003). (c) Number of days per century with one or more sigtors 1921–95 (Concannon et al. 2000). (d) Count of all tornado events, 2003–11 (Smith et al. 2012). (e) Number of sigtors 1950–2004 (Ashley 2007). (f) Number of days with sigtor per century, 1961–2010 (Carbin et al. 2012).
Citation: Weather and Forecasting 29, 2; 10.1175/WAF-D-13-00057.1
There are several issues that have not been fully addressed in prior tornado climatology studies. First of all, the Fujita scale was not published until 1971 (Fujita 1971; Edwards et al. 2010), and was not adopted for rating tornadoes in their near-immediate aftermath in the United States until 1973. So, tornadoes before this time were rated years after the fact in the SPC database, using newspaper stories and photographs that likely overemphasized the damage, causing overrating of some tornadoes prior to 1973 (e.g., Schaefer and Edwards 1999; Brooks and Craven 2002; Schaefer and Schneider 2002; Anderson et al. 2007). Also, there has been a steady increase in annual reported tornadoes in the United States, even since 1973, due to better spotting techniques, National Weather Service (NWS) warning verification procedures, and improved detection by radar. This overall increase has primarily been due to an increase in weak tornadoes (F0 or F1, hereinafter referred to as weaktors). However, the reporting of significant tornadoes (F2 or greater, hereafter referred to as sigtors) has remained relatively stable since 1973. Additionally, as seen from the results of various researchers in Fig. 1, most tornado climatologies have used either tornado days or the number of tornadoes as the basis for their studies, not utilizing pathlength. However, pathlength seems more desirable because it is more proportional to the area impacted by tornadoes than tornado numbers or tornado days.
The purpose of this research is to further our understanding of the regions of greatest tornado risk in the United States by addressing the above issues in an objective analysis. Section 2 will discuss the data and methodology used in the study. Section 3 will examine some trends in tornado reporting and discuss the issues mentioned above, including the importance of pathlength in assessing tornado risk. This information will help remove some biases that are present in other studies, producing an objective analysis of tornado risk in the United States in section 4. Section 5 contains a summary and conclusions.
2. Data collection and methodology
The primary data source for this study is the SPC tornado database, which provides the date and time, location, intensity [Fujita or enhanced Fujita (EF) scale], pathlength, and other statistics on nearly every tornado in the United States for 1950–2011. We focus primarily on data from 1973 through 2011, to eliminate any bias from tornadoes rated after the fact. We also use only significant tornadoes (F2 or greater and EF2 or greater since the enhanced Fujita scale was put in place), as done by many other authors (e.g., Concannon et al. 2000; Meyer et al. 2002; Ashley 2007; Carbin et al. 2012). This is due to the fact that the number of reported sigtors has remained fairly steady since 1973, implying that factors improving tornado reporting efficiency over time have not affected the reporting of sigtors nearly as much as they have affected reporting of weaktors. As pointed out by Britt and Glass (2013), this also removes most tropical cyclone–related tornadoes, which tend to be weak [EF0 or EF1; e.g., Schultz and Cecil (2009)]. Maps showing the smoothed average annual pathlength of tornadoes within 40 km (25 mi) of a point were generated using kernel density estimation (KDE, e.g., O’Sullivan and Unwin 2003).
KDE is a very common method used in spatial analysis and has been applied to tornado risk in several previous papers (e.g., Brooks et al. 2003; Dixon et al. 2011; Smith et al. 2012; Marsh and Brooks 2012). As discussed in detail by Dixon et al. (2011), KDE analysis implies that spatial patterns have magnitudes at every given point, as opposed to only the places directly affected (in this case, by tornadoes) (e.g., O’Sullivan and Unwin 2003). The KDE method used in this study employs the Epanechnikov quadratic kernel probability density function as calculated by ESRI ArcGIS Spatial Analyst software (Silverman 1986; de Smith et al. 2007; Dixon et al. 2011; Marsh and Brooks 2012; Smith et al. 2012). There are several other types (i.e., shapes) of kernel functions, but Dixon and Mercer (2012) shows that the differences in spatial patterns due to kernel shape are negligible compared to those related to the radius of the kernel, which is consistent with previous research (Silverman 1986; de Smith et al. 2007). Throughout this study, a kernel radius of 250 km is used because of the relatively short study period of the current analysis. The steps we used in performing the KDE analysis were as follow: 1) The 250-km probability density function (KDE) was applied to each tornado path on a 1-km2 grid. This results in each 1-km2 grid cell within 250 km of the tornado path in question receiving a portion of the total tornado path (according to its distance from the path). 2) The sum of each point’s fractional pathlengths (measured in kilometers of tornado path per square kilometer) for all tornadoes to affect that grid cell was calculated on a 5-km output grid. So, while only 1 out of every 25 grids is displayed, this coarser output grid changes the resolution of the map and does not change the numerical results (Dixon et al. 2011). 3) Every cell is then multiplied by the approximate area of a 40-km-radius circle (5024 km2) to represent a meaningful risk metric (in kilometers of tornado path per 5024 km2, or kilometers of tornado path within 40 km of the cell), and to be consistent with SPC’s tornado risk forecasts (Kay and Brooks 2000).
3. Analysis of trends and selection of dataset
a. Changes in tornado reporting efficiency since 1950
Since 1950, according to the SPC tornado database, there has been a large increase in the number of reported tornadoes per year in the United States (Fig. 2a). The increase has been fairly steady (linear correlation coefficient of 0.85) over the 62-yr reporting period, with an increase of about 16 tornadoes per year reported. Only slightly improved is the third-order polynomial fit (correlation coefficient of 0.86), which highlights the rapid increase in reported tornadoes from 1950 through 1965, followed by a slower rate of growth in the 1970s and 1980s, and then accelerated growth in reported tornadoes after 1990 [during the Next Generation Weather Radar (NEXRAD) era].
(a) All reported tornadoes and (b) significant tornadoes by year in the United States for 1950–2011. In each chart, the dark curves indicate the actual number of reported tornadoes, blue dashed lines indicate the best linear-curve fit, and the orange curve indicates the best third-order-polynomial fit. In (b), the periods for 1950–73 and 1973–2011 are separated to account for the beginning of the near-real-time rating of tornadoes, as opposed to after the fact rating.
Citation: Weather and Forecasting 29, 2; 10.1175/WAF-D-13-00057.1
Most of this overall increase is not due to an increase in tornado occurrence but to increases in tornado reporting (e.g., Kunkel et al. 2013). This increase in tornado reporting efficiency has likely been due to better spotting techniques with time (e.g., Coleman et al. 2011), NWS warning verification procedures that prompted more frequent storm surveys (McCarthy and Schaefer 2004), and the implementation of the Weather Surveillance Radar-1988 Doppler (WSR-88D) network in the early 1990s (McCarthy and Schaefer 2004). This radar system alerts the NWS to areas with possible tornadoes (including weaktors), even in unpopulated areas. The NWS then conducts storm surveys, finding tornadoes they may not have found before the WSR-88D network came into being.
If only sigtors (F2 or greater) are considered, there is practically no overall increase since 1950 (Fig. 2b), and there is actually a decrease in sigtors since 1973. The number of reported sigtors increased during the 1950s as the modern era of tornado forecasting and warning began (e.g., Schaefer 1986; Doswell et al. 1999). After 1973, when tornadoes began to be rated in near-real-time poststorm surveys using the Fujita scale, the average annual number of sigtors decreased dramatically, from 219 during the period 1950–72, to 165 during the period 1973–2011. Results of our analyses (Fig. 2) support the assertions of Schaefer and Edwards (1999) that tornadoes prior to 1973 were likely overrated in many cases. Therefore, our remaining analyses are focused on tornado data since 1973.
b. Tornado reports 1973–2011: Sigtors contrast with weaktors
During the period from 1973 through 2011, the total number of annually reported tornadoes in the United States continued to increase, but the number of sigtors has remained fairly steady (Fig. 2). To more specifically examine the geographic differences around this issue, Fig. 3 shows the yearly number of weaktors and sigtors per 1000 km2, comparing the Great Plains with the southeastern United States. In this paper, the Great Plains (GP) is considered to be the states of Texas, Oklahoma, Kansas, Nebraska, and Iowa, the traditional tornado alley (covering 1.4 × 106 km2). The southeastern United States includes Arkansas, Louisiana, Mississippi, Tennessee, and Alabama (covering 642 000 km2).
(a),(b) Plot of annual numbers of weak tornadoes (F0 and F1) per 1000 km2 of land area in the GP and southeastern United States, respectively. (c),(d) As in (a),(b), but for significant tornadoes (F2–F5).
Citation: Weather and Forecasting 29, 2; 10.1175/WAF-D-13-00057.1
Note that Fig. 3a shows that the number of reported weaktors in the GP continued to increase from 1973 through about 1991, before reaching relative stability. This is likely due, at least partially, to the large increase in research-related and nonprofessional storm chasing in that region starting in the 1980s and growing rapidly in the 1990s. However, in the southeastern United States, storm spotting is more difficult. In many areas, terrain is more rugged, and much more of the area is covered with trees than in the GP. In addition, a larger percentage of southeastern United States tornadoes occur at night. Using a sun zenith-angle algorithm to determine sunrise and sunset by date and latitude, it was determined that 48.6% of all southeastern U.S. tornadoes between 1973 and 2011 occurred at night, compared to 39.3% in the GP. Therefore, the number of reported weaktors in the southeastern United States continued along an upward trend (Fig. 3b) even after 2000, unlike the GP, as advances in technology improved reporting capabilities.
It is reasonable to assume that, due to their larger size and longer duration, sigtors are less likely to go unreported (e.g., Concannon et al. 2000; Brooks 2004). According to the SPC tornado database (1973–2011), the average pathlength of a weaktor was 2.9 km, and the average path area (length × width) was only 0.33 km2. By contrast, the average pathlength of a sigtor was 15.1 km, and the average path area was 6.26 km2, about 19 times as large as weaktors. This makes significant tornadoes much more likely to be detected, as their damage is more severe and they affect a much larger area. Anderson et al. (2007) found that, in most areas, the probability of detection of sigtors in rural areas is greater than that for weaktors, consistent with the above analysis. Also, the number of sigtors in both regions has been steady or even slowly declining since 1973 (Figs. 3c and 3d). The slight decline might be due to more professional damage evaluations due to the NWS warning verification program and its growth during the 1980s. However, using the linear fit to each chart, the magnitude of the change is still much larger with weaktors than it is with sigtors. The number of reported weaktors, combining data from both the GP and the southeastern United States, is increasing at the rate of about 6 tornadoes per 106 km2 yr−1, while the number of sigtors is decreasing at a rate of 0.7 tornadoes per 106 km2 yr−1.
In summary, even during the “modern era” (1973–2011) the number of reported weaktors has changed by nearly an order of magnitude more than the number of sigtors. Also, there is a greater likelihood for the reporting of sigtors as opposed to weaktors due to their longer pathlengths, larger path areas, greater damage, and other factors discussed above. Their much more stable reporting pattern since 1973 and their higher likelihood of being reported consistently over time make sigtors the best dataset to use when analyzing the areas with the greatest tornado risk in the United States.
c. Pathlength versus tornado count–tornado days
As illustrated by Dixon and Mercer (2012), large differences in spatial patterns can occur if one uses point analysis [i.e., the point of tornadogenesis; e.g., Brooks et al. (2003)] instead of pathlength analysis (e.g., Dixon et al. 2011). Given that the pathlength of a tornado is much more proportional to the area it affects, its destructive capability, and its overall impact on the risk of a tornado striking an area within its radius of influence in the KDE, pathlength analysis is used in this study unless otherwise noted. In this study, those events in the database that lacked a termination point or showed identical initiation and termination points were altered by adding a termination point 100 m due north of the initiation. The ultimate justification for this method is to reduce underestimation of the spatial risk in areas that often experience short-lived and/or slow-moving tornadoes, as even a short path is significantly more impactful than the unrealistic point locations listed in the database.
Some authors use tornado days as their measure of tornado risk, to minimize the temporal trends in reporting frequency that we have already addressed. However, tornado days are not the best measure of tornado risk, since one single F0 tornado occurring in the vicinity of a point would have the same input to the climatology as an outbreak of 25 tornadoes, some of them significant, in the same area on one day. This study makes use of a relatively stable set of data (1973–2011 significant tornadoes), and Dixon et al. (2011) show only minor variations in the spatial patterns of pathlength analyses of tornado days and total tornadoes, so all significant tornado events will be analyzed. Further, use of all sigtors allows for a clearer understanding of the climatological differences between the GP and southeastern United States, as shown by a quick analysis of the destruction potential index (DPI; Thompson and Vescio 1998; Doswell et al. 2006). DPI is defined as the area covered by the tornado (pathlength × path width) multiplied by the Fujita or enhanced Fujita rating plus one. Because it combines the area covered with the intensity, the DPI for a given tornado provides a reasonable measurement of that tornado’s potential impact on society. Since 1973, the median DPI per tornado day is 0.18 in the GP, but 0.45 in the southeastern United States. Therefore, using tornado days greatly lessens the modeled impact of tornadoes in the southeastern United States relative to the GP.
4. Objective analysis of tornado risk
a. Initial analysis
Given that the intensity data on tornadoes greatly improved starting in 1973, and sigtors provide the most stable dataset for analysis of tornado risk, we first examine sigtors over the United States from 1973 through 2011 (Fig. 4). In terms of the average annual sum of the pathlengths of sigtors passing within 40 km (25 mi) of a point, there is a broad area extending from much of the classic tornado alley in the GP eastward into the Midwest and southeastern United States that experiences >3 km of annual pathlength. A greater risk area (>5 km of annual pathlength) extends primarily from Oklahoma to northwest Georgia, with isolated areas in Iowa and near the Nebraska–Kansas border. The maximum risk area (>7 km of annual pathlength) covers the eastern part of the primary risk area, from central Arkansas into central Mississippi and northern Alabama. A small region of even higher risk (>9 km of annual pathlength) extends from near Jackson, Mississippi, to Huntsville and Birmingham, Alabama. This distribution is quite consistent with that shown in the much smaller dataset by Smith et al. (2012; see Fig. 1d), and fairly consistent with the larger dataset shown by Carbin et al. (2012; see Fig. 1f). Initially, Fig. 4 indicates that the area of greatest tornado risk in the United States extends from Oklahoma to Alabama and Tennessee, with the greatest risk extending from eastern Arkansas into central Mississippi and northern Alabama.
KDE analysis showing the average annual pathlength (km) of all significant (F2–F5) tornadoes passing within 40 km (25 mi) of a point between 1973 and 2011.
Citation: Weather and Forecasting 29, 2; 10.1175/WAF-D-13-00057.1
To determine the impact of the single, very large outbreak of long-track significant tornadoes on 27 April 2011 (e.g., Knupp et al. 2014) across parts of the southeastern United States (Fig. 5), Fig. 6 shows a KDE of significant tornadoes during 1973–2010, excluding the events of 2011. The high risk area (>5 km annual pathlength) still extends from central Oklahoma to northwest Georgia, with small areas of high risk near the Nebraska–Kansas border, in central Iowa, and southern Indiana. The highest risk area (>9 km of annual pathlength) covers eastern Arkansas, central Mississippi, and northern Alabama, and it seems that the region of maximum tornado risk indicated in Fig. 4, from Oklahoma to Alabama and Tennessee, was already apparent in the data prior to the 2011 tornadoes.
Plot of all significant tornado tracks (and tornado F-scale intensities) from 0600 UTC 27 Apr to 0600 UTC 28 Apr 2011 over the southeastern United States. (Plot constructed using SPC Severe Plot 3.0.)
Citation: Weather and Forecasting 29, 2; 10.1175/WAF-D-13-00057.1
As in Fig. 4, but between 1973 and 2010.
Citation: Weather and Forecasting 29, 2; 10.1175/WAF-D-13-00057.1
If one examines the average annual DPI of tornadoes, the same general area is once again highlighted, as shown in the KDE analysis for DPI (Fig. 7), except there is a more prominent maximum centered over eastern Mississippi and western Alabama. Even if one only considers the number of sigtors, KDE analysis of starting points of sigtors (without any input from pathlength; Fig. 8) shows that the region from Oklahoma to Alabama is also the region with the most individual sigtors, but the maximum over Mississippi and Alabama shown in Fig. 4 is not as apparent. This is likely due to the fact that the average pathlength for a sigtor in the southeastern United States is 18.3 km (where many more tornadoes occur during the cool season and move quickly, covering more ground), significantly longer than the average pathlength for a sigtor in the GP (14.8 km).
As in Fig. 4, but showing the average annual total DPI (“Fujita” km2).
Citation: Weather and Forecasting 29, 2; 10.1175/WAF-D-13-00057.1
As in Fig. 4, but showing the average annual number of significant (F2–F5) tornado-initiation locations.
Citation: Weather and Forecasting 29, 2; 10.1175/WAF-D-13-00057.1
b. Lack of bias due to underrated tornadoes in the GP
An argument can be made that some of the tornadoes in the GP do not make it into our calculations because, even though they had the wind speeds of a sigtor, they were rated as weaktors because there was simply nothing for them to destroy in the thinly populated open grasslands and fields in some parts of the GP (e.g., Doswell and Burgess 1988). However, we present three points that suggest any such bias does not fundamentally change the spatial patterns of tornado risk in the United States.
NEXRAD radars went online nationwide in the early 1990s, making the detection of all tornadoes, including weak ones, more likely. Also since the early 1990s, with the Verification of the Origins of Rotation in Tornadoes Experiment (VORTEX; Rasmussen et al. 1994) and the movie Twister, tornado chasing has become almost ubiquitous during days with tornado risk, especially in the GP. This has also reduced the likelihood of unreported tornadoes. Furthermore, trend data have shown that annual weaktors have become fairly steady over the GP since 1992. A KDE analysis of 1992–2011 tornado pathlength was performed (not shown). Even if one assumes that 10% of all weaktor pathlength over eastern Kansas and eastern Nebraska was associated with underreported sigtors, adding that pathlength to the reported sigtor pathlength still indicates that those areas had less tornado risk than most of Oklahoma, and much lower risk than parts of Alabama, Arkansas, and Mississippi.
Because the average pathlength of a sigtor is more than 5 times that of a weaktor, the presence of many long-track weaktors in the GP could indicate an underrating of some sigtors there. A KDE analysis of 1973–2011 tornadoes with pathlengths greater than 16.1 km (10 mi, close to the average pathlength of a sigtor) shows a very similar pattern to that of sigtors (Fig. 9). An area of moderate long-track tornado risk does extend through eastern Kansas, southeastern Nebraska, and Iowa. This may indicate that some tornadoes in these states are underrated and, therefore, not included in the sigtor database. However, many of these areas have population densities not that different than those is much of Arkansas and Mississippi (Fig. 10), so it is likely that most sigtors with pathlengths greater than 10 mi would be similarly likely to damage something and be classified as a sigtor.
The detection frequency for weaktors has increased since 1992, especially in the southeastern United States, due to factors mentioned in section 3a. By applying a linear curve fit to the number of weaktors during 1992–2011, the rate of increase in detection frequency may be calculated, and factors applied to each year to simulate a constant detection frequency (assuming the average annual number of actual weaktors was constant from 1992 through 2011). These calculations were performed for both the southeastern United States and the GP, and factors were applied to all weaktors by year in each region to force the detection frequency to be approximately constant. Using the detrended number of weaktors calculated above, and the reported number of sigtors (since sigtors show almost no trend), 8.8% of all tornadoes in the GP from 1992 through 2011 were sigtors, according to the SPC database. Using similar calculations, 12.8% of all tornadoes in the southeastern United States during that time period were sigtors. Making the assumption that the 4% difference is due to underreporting of sigtors in the GP (i.e., each region actually has the same ratio of sigtors to all tornadoes), a number of 1992–2011 tornadoes in the GP that were officially rated as weaktors in the SPC tornado database were artificially increased to sigtor status. In addition to sigtors, any weaktor that occurred in the GP with a path width greater than 249 yd (1 yd = 0.9144 m) was included as a sigtor in a new database of tornadoes from 1992 through 2011. Only officially rated sigtors in the southeastern United States and the rest of the country were included in the database. This way, both the southeastern United States and the GP had a sigtor ratio near 12.8%. A KDE analysis, similar to Fig. 4, was then performed on this database with the added sigtors in the GP (Fig. 11). Since Fig. 11 is quite similar to Fig. 4, it is apparent that underreporting of sigtors in the GP is not a major factor in the KDE analysis of sigtors.
KDE analysis of tornadoes (1973–2011) with pathlengths > 10 mi.
Citation: Weather and Forecasting 29, 2; 10.1175/WAF-D-13-00057.1
Population density by county in the United States, 2011 (U.S. Census Bureau 2013).
Citation: Weather and Forecasting 29, 2; 10.1175/WAF-D-13-00057.1
As in Fig. 4, but between 1992 and 2011 and using the lowered criterion for significant tornadoes in the GP (see text).
Citation: Weather and Forecasting 29, 2; 10.1175/WAF-D-13-00057.1
5. Summary and conclusions
Only data since 1973 were used in this study because many tornadoes were overrated before 1973 (section 3a). Only significant tornadoes were included because they are much more likely to be reported and their frequencies have been fairly steady with time since 1973, unlike weak tornadoes (section 3b). Finally, pathlengths were used in KDE analysis as opposed to tornadogenesis points or tornado days. Our reasoning for favoring the use of tornado pathlengths (representing tornado frequency and severity) as opposed to points of tornadogenesis or tornado days (representing only tornado frequency) is explained in section 3c. Given the SPC tornado database of sigtors occurring from 1973 through 2011, Fig. 4 shows that the region with the greatest risk of tornadoes in the United States is a roughly west–east-oriented area, from central Oklahoma, through Arkansas and northern Louisiana, western and middle Tennessee, most of Mississippi, and northern and central Alabama. Even upon examination of tornado count, our analysis of the region with the highest tornado risk does not change appreciably. There is still risk for tornadoes in many other areas, especially across parts of Kansas, Nebraska, Iowa, Illinois, Indiana, and Kentucky. No state is totally immune to tornadoes. This study points out that the greatest risk is in the southern United States.
The seemingly legitimate argument that lower population density in the GP causes the underrating of tornadoes, and therefore some sigtors to be rated as weaktors, is discussed and mostly discounted in section 4c on the basis of changing some weaktors to sigtors in eastern Kansas and eastern Nebraska, and distribution of moderate- to long-track tornadoes. In addition, when reporting trends are mathematically removed for weaktors and many weaktors in the GP are artificially inflated to sigtor status, forcing the ratio of sigtors to all tornadoes to be the same in the southeastern United States and the GP, the same general region of risk shown in Fig. 4 is highlighted. Also, the destruction potential index (DPI) of tornadoes in the southeastern United States is much higher, on average. The DPI for a given tornado provides a reasonable measurement of a tornado’s potential impact on society, and the overall risk to people due to a tornado. A KDE analysis for DPI was performed (Fig. 7), and the region of highest tornado risk from Oklahoma to Alabama is confirmed, with a significant maximum extending from central Mississippi into northern Alabama.
The tornado risk area outlined in this paper is consistent with a trend in findings by several authors in recent papers who only examine sigtors, yet use a large dataset covering at least 30 yr (e.g., Concannon et al. 2000; see Fig. 1c; Ashley 2007; see Fig. 1e; Carbin et al. 2012; see Fig. 1f). Therefore, we present it as the part of the United States with the highest risk of tornadoes. This area is so clearly different than the tornado alley discussed by the media for many years that many Americans are totally unaware of the tornado risk in some of the areas outlined in this study.
Acknowledgments
This research was funded by the National Science Foundation (NSF Award AGS-1110622). Reviews of the manuscript by Jon Davies, Greg Forbes, Kevin Knupp, Ryan Wade, and Todd Murphy were very helpful and improved the manuscript greatly.
REFERENCES
Anderson, C. J., Wikle C. K. , Zhou Q. , and Royle J. A. , 2007: Population influences on tornado reports in the United States. Wea. Forecasting, 22, 571–579.
Ashley, W. S., 2007: Spatial and temporal analysis of tornado fatalities in the United States: 1880–2005. Wea. Forecasting, 22, 1214–1228.
Britt, M. F., and Glass F. H. , cited 2013: The tornado climatology of the St. Louis Weather Forecast Office County Warning Area. National Weather Service Weather Forecast Office, St. Louis, MO. [Available online at http://www.crh.noaa.gov/lsx/?n=torcli_data.]
Brooks, H. E., 2004: On the relationship of tornado path length and width to intensity. Wea. Forecasting, 19, 310–319.
Brooks, H. E., and Craven J. P. , 2002: A database of proximity soundings for significant severe thunderstorms, 1957–1993. Preprints, 21st Conf. on Severe Local Storms, San Antonio, TX, Amer. Meteor. Soc., 16.2. [Available online at https://ams.confex.com/ams/pdfpapers/46680.pdf.]
Brooks, H. E., Doswell C. A. III, and Kay M. P. , 2003: Climatological estimates of local daily tornado probability for the United States. Wea. Forecasting, 18, 626–640.
Carbin, G., Heinselman P. , and Stensrud D. , 2012: Current challenges in tornado forecast and warning. Workshop on Weather Ready Nation: Science Imperatives for Severe Thunderstorm Research, M. K. Lindell and H. Brooks, Eds., NOAA–NSF, A-9–A-16.
Coleman, T. A., Knupp K. R. , Spann J. , Elliott J. B. , and Peters B. E. , 2011: The history (and future) of tornado warning dissemination in the United States. Bull. Amer. Meteor. Soc., 92, 567–582.
Concannon, P. R., Brooks H. E. , and Doswell C. A. III, 2000: Climatological risk of strong to violent tornadoes in the United States. Preprints, Second Symp. on Environmental Applications, Long Beach, CA, Amer. Meteor. Soc., 212–219.
de Smith, M. J., Goodchild M. F. , and Longley P. A. , 2007: Geospatial Analysis: A Comprehensive Guide to Principles, Techniques and Software Tools. 3rd ed. Troubador Publishing, 516 pp.
Dixon, P. G., and Mercer A. E. , 2012: Reply to “Comments on ‘Tornado risk analysis: Is Dixie Alley an extension of Tornado Alley?’” Bull. Amer. Meteor. Soc., 93, 408–410.
Dixon, P. G., Mercer A. E. , Choi J. , and Allen J. S. , 2011: Tornado risk analysis: Is Dixie Alley an extension of Tornado Alley? Bull. Amer. Meteor. Soc., 92, 433–441.
Doswell, C. A., III, and Burgess D. W. , 1988: On some issues of United States tornado climatology. Mon. Wea. Rev., 116, 495–501.
Doswell, C. A., III, Moller A. R. , and Brooks H. E. , 1999: Storm spotting and public awareness since the first tornado forecasts of 1948. Wea. Forecasting, 14, 544–557.
Doswell, C. A., III, Edwards R. , Thompson R. L. , Hart J. A. , and Crosbie K. C. , 2006: A simple and flexible method for ranking severe weather events. Wea. Forecasting, 21, 939–951.
Edwards, R., LaDue J. G. , Ferree J. T. , Scharfenberg K. , Maier C. , and Coulbourne W. L. , 2010: The enhanced Fujita scale: Past, present, and future. Preprints, 25th Conf. on Severe Local Storms, Denver, CO, Amer. Meteor. Soc., 4A.1. [Available online at https://ams.confex.com/ams/pdfpapers/175400.pdf.]
Fujita, T. T., 1971: Proposed characterization of tornadoes and hurricanes by area and intensity. SMRP Research Paper 91, University of Chicago, Chicago, IL, 42 pp. [Available from Wind Engineering Research Center, Box 41023, Lubbock, TX 79409.]
Gagan, J. P., Gerard A. , and Gordon J. , 2010: A historical and statistical comparison of “Tornado Alley” to “Dixie Alley.” Natl. Wea. Dig., 34 (2), 145–155.
Kay, M. P., and Brooks H. E. , 2000: Verification of probabilistic severe storm forecasts at the SPC. Preprints, 20th Conf. on Severe Local Storms, Orlando, FL, Amer. Meteor. Soc., 9.3. [Available online at https://ams.confex.com/ams/Sept2000/techprogram/paper_15921.htm.]
Kellogg, B., and Forbes G. S. , cited 2013: Top 10 tornado cities. The Weather Channel. [Available online at http://www.weather.com/news/tornado-central/top-tornado-cities-2013-20130415.]
Kelly, D. L., Schaefer J. T. , McNulty R. P. , Doswell C. A. III, and Abbey R. F. , 1978: An augmented tornado climatology. Mon. Wea. Rev., 106, 1172–1183.
Knupp, K. R., and Coauthors, 2014: Meteorological overview of the devastating 27 April 2011 tornado outbreak. Bull. Amer. Meteor. Soc., in press.
Kunkel, K. E., and Coauthors, 2013: Monitoring and understanding trends in extreme storms: State of knowledge. Bull. Amer. Meteor. Soc., 94, 499–514.
Marsh, P. T., and Brooks H. E. , 2012: Comments on “Tornado risk analysis: Is Dixie Alley an extension of Tornado Alley?” Bull. Amer. Meteor. Soc., 93, 405–407.
McCarthy, D. W., and Schaefer J. T. , 2004: Tornado trends over the past thirty years. Preprints, 14th Conf. on Applied Climatology, Seattle, WA, Amer. Meteor. Soc., 3.4. [Available online at https://ams.confex.com/ams/84Annual/techprogram/paper_72089.htm.]
McNulty, R. P., 1995: Severe and convective weather: A Central Region forecasting challenge. Wea. Forecasting, 10, 187–202.
Meyer, C. L., Brooks H. E. , and Kay M. P. , 2002: A hazard model for tornado occurrence in the United States. Preprints, 16th Conf. on Probability and Statistics in the Atmospheric Sciences/13th Symp. on Global Change and Climate Variations, Orlando, FL, Amer. Meteor. Soc., J3.6. [Available online at https://ams.confex.com/ams/pdfpapers/27595.pdf.]
O’Sullivan, D., and Unwin D. J. , 2003: Geographic Information Analysis. J. Wiley and Sons, 436 pp.
Rasmussen, E. N., Straka J. M. , Davies-Jones R. , Doswell C. A. III, Carr F. H. , Eilts M. D. , and MacGorman D. R. , 1994: Verification of the Origins of Rotation in Tornadoes Experiment: VORTEX. Bull. Amer. Meteor. Soc., 75, 995–1006.
Schaefer, J. T., 1986: Severe thunderstorm forecasting: A historical perspective. Wea. Forecasting, 1, 164–189.
Schaefer, J. T., and Edwards R. , 1999: The SPC tornado/severe thunderstorm database. Preprints, 11th Conf. on Applied Climatology, Dallas, TX, Amer. Meteor. Soc., 6.11. [Available online at https://ams.confex.com/ams/99annual/abstracts/1360.htm.]
Schaefer, J. T., and Schneider R. S. , 2002: The robustness of tornado hazard estimates. Preprints, Third Symp. on Environmental Applications, Orlando, FL, Amer. Meteor. Soc., 4.2. [Available online at https://ams.confex.com/ams/pdfpapers/27694.pdf.]
Schaefer, J. T., Kelly D. L. , Doswell C. A. , Galway J. G. , Williams R. J. , McNulty R. P. , Lemon L. R. , and Lambert B. D. , 1980: Tornadoes, when, where, how often. Weatherwise, 33, 52–59.
Schultz, L. A., and Cecil D. J. , 2009: Tropical cyclone tornadoes, 1950–2007. Mon. Wea. Rev., 137, 3471–3484.
Silverman, B. W., 1986: Density Estimation for Statistics and Data Analysis. Chapman and Hall, 175 pp.
Smith, B. T., Thompson R. L. , Grams J. S. , Broyles C. , and Brooks H. E. , 2012: Convective modes for significant severe thunderstorms in the contiguous United States. Part I: Storm classification and climatology. Wea. Forecasting, 27, 1114–1135.
Thompson, R. L., and Vescio M. D. , 1998: The destruction potential index—A method for comparing tornado days. Preprints, 19th Conf. on Severe Local Storms, Minneapolis, MN, Amer. Meteor. Soc., 280–282.
U.S. Census Bureau, cited2013: Population density for counties and Puerto Rico municipios: July 1, 2011. [Available online at https://www.census.gov/popest/data/maps/2011/PopDensity_11.pdf.]
