1. Introduction
“Whole frame houses tossed off foundations; steel-reinforced concrete structures badly damaged; automobile-sized missiles generated; incredible phenomena can occur.”
The F scale was quickly adopted by the National Weather Service (NWS) in the years subsequent to its development, with an extensive effort undertaken by Grazulis (1993) to create a climatology of significant (F2 or greater) tornadoes with rating estimates dating back to approximately 1880.
Not long after the F scale was implemented, questions were raised about the wind speed estimates, particularly in stronger tornadoes. Minor et al. (1977) synthesized information from engineering studies that indicated that no evidence existed of wind speeds in tornadoes exceeding 112 m s−1 (250 mph) and speculated that tornado wind speeds are unlikely to exceed 112–123 m s−1 (250–275 mph). In response to concerns about the overestimation of wind speeds at higher F-scale ratings and the lack of consideration for integrity of structures in the original F scale, Fujita (1992) acknowledged that structural integrity should be taken into account when assigning an F-scale rating and created a table to adjust for the raw damage-based ratings to buildings to the final F-scale rating based on their construction. The F5 tornadoes at Jarrell, Texas, on 27 May 1997; Jefferson County, Alabama, on 8 April 1998; and Moore, Oklahoma, on 3 May 1999 reignited discussions about the likelihood of overestimation of wind speeds necessary to cause damage at the higher ratings of the F scale. Phan and Simiu (1998) suggested that the extreme damage noted in Jarrell could have been attributed to winds in the F3 range, given the long duration of tornado wind exposure at any given point and the quality of attachment of homes to their foundations within the Jarrell track. Mehta and Carter (1999) found that the damage in Jefferson County could have been produced by wind speeds no greater than 51–67 m s−1 (114–150 mph). Marshall (2002) determined that structural failures in Moore likely occurred at wind speeds in the F2 range.
Another key event to impact the climatology of F5 ratings during the F-scale era was the La Plata, Maryland, tornado of 28 April 2002 (U.S. Department of Commerce 2002). The La Plata tornado was rated F4 on an initial survey by NWS Baltimore/Washington, prior to being upgraded to F5 during subsequent analysis. However, a closer investigation by a structural engineering expert revealed that the structures rated F5 all featured significant structural flaws, many of which were homes that were poorly attached or unattached to their foundations. The La Plata tornado was ultimately downgraded back to F4 in the final analysis. The combination of engineering estimates of the wind speeds necessary to cause intense tornado damage, the likelihood that the F scale substantially overestimated those wind speeds, and the evolution of the La Plata tornado rating likely led to the decrease in F4–F5 tornado ratings from 2000 to 2006 noted in Edwards et al. (2021).
2. The F- to EF-scale transition
The enhanced Fujita scale (EF) was developed in 2006 to address concerns about the overestimation of wind speeds in stronger tornadoes and provided a more complete catalog of damage indicators (DIs) to estimate tornado wind speeds from damage [Wind Science and Engineering Center (WSEC) 2006]. Unlike the F scale, the EF scale was developed as a damage scale, from which wind speeds were then estimated. The EF scale features 28 different DIs, each of which has a list of different degrees of damage (DODs) each DI may exhibit. However, of these 28 DIs, only 12 have DODs with wind speed ranges extending into the EF5 category. And of those 12 DIs that extend into the EF5 range, only four have DODs where the expected value wind speed falls into the EF5 range: DI 11 (large shopping malls), DI 18 (midrise buildings of 5–20 stories), DI 19 (high-rise buildings exceeding 20 stories), and DI 20 (institutional buildings; Fig. 1). There is no direct quantitative relationship between the EF ratings and wind speeds—the wind speed ranges were derived from a relationship between F-scale and EF-scale estimates of wind speeds needed to cause damage to preserve consistency in the ratings between the two scales.
Figure 1 from Lyza et al. (2024) illustrating the wind speed ranges for each DI of the EF scale. Bold bars indicate the range from the expected value wind gust speed of the lowest DOD to the expected value of the highest DOD, while the thin tails represent the range from the lower-bound wind gust speed of the lowest DOD to the upper-bound wind gust speed of the highest DOD. ©American Meteorological Society. Used with permission.
Citation: Bulletin of the American Meteorological Society 106, 8; 10.1175/BAMS-D-24-0066.1
Notably, the expected value for the top DOD for DI 2 (one- and two-family residences) features an expected wind gust value of 200 mph, the upper bound of the EF4 wind speed range. As noted in Edwards et al. (2021), well-built single-family homes constitute the “common DI” between the F and EF scales. In the F scale, a well-built home being swept clean from its foundation constituted an F5 rating; however, in the EF scale, the expected wind gust value for a well-built home being swept away is 200 mph or the upper bound of EF4. Intriguingly, this discrepancy is a by-product of how the final EF-scale wind speed ranges were determined. The raw EF5 wind speed range described in WSEC (2006), which was derived from a linear regression relationship between the individual DI/DOD wind speeds determined through the “expert elicitation” process for the EF scale as a function of independent legacy F-scale wind speed estimations, was 200–234 mph. However, after this linear regression derivation, the committee rounded each wind speed range to 5-mph increments, with the rounding applied at the top of each range. Therefore, the raw 168–199-mph EF4 wind speed range was adjusted to 166–200 mph, which led to the starting threshold for EF5 being changed from 200 to 201 mph (WSEC 2006, their Table 6). Consequently, under the strictest application of the EF scale, to attain an EF5 rating from a single-family home being swept off its foundation, the home must technically be built above building code, which is a fundamental break from the F scale and will inherently reduce the number of EF5 DIs found in surveys. It is worth noting, however, that the basis for what is considered “well-built” or “built to code” is ever evolving. For instance, the F5 example picture in Fujita (1971), which featured a house built on a concrete masonry unit foundation with little obvious sign of attachment, would be unlikely to meet the expected construction quality of a single-family home as described in the EF scale (WSEC 2006), which explicitly describes the need for structures to maintain a consistently strong load path from the exterior to the foundation.
Upon the initial implementation of the EF scale, however, EF5 ratings were given to several tornadoes on the basis of damage to single-family homes without directly noting that these homes were built at standard construction codes (Table 1). In the 20 May 2013 Moore, Oklahoma, tornado, Burgess et al. (2014) developed a rubric for determining EF5 damage to homes based on characteristics of the foundation anchoring and estimation of how much force was applied to the foundation anchors before the house completely failed (i.e., how long did the walls remain attached prior to failure and removal from the foundation, as indicated by the degree of anchor bolt bending, bolt spacing, presence of proper washers and nuts on anchor bolts, and fraction of sill plates remaining). Using this rubric, they identified nine homes that sustained EF5 damage in Moore in 2013.
Summary of all EF5-rated tornadoes in the United States and the justification for their ratings.
In the Parkersburg, Iowa, tornado of 25 May 2008 and the Hackleburg, Alabama, tornado of 27 April 2011, surveyors noted additional contextual evidence beyond the sweeping of homes in applying EF5 ratings. In the Hackleburg tornado, the lofting of vehicles 150–200 yards and the “wind rowing” of debris, or collection of debris in long, narrow rows downstream of its source, were used as evidence of EF5 damage in the city of Hackleburg (NWS Birmingham 2023). In fact, while dozens of homes were swept away in total along the 164-km (102 mi) track (Lyza et al. 2022), only one home was explicitly noted to have actually been anchored to its foundation (NWS 2023). In the Parkersburg tornado, Marshall et al. (2008b) explicitly noted that homes being swept away were insufficient to apply the EF5 rating, but that an EF5 rating was applied based on a combination of home damage and extreme granularization of debris and ground scouring along the core of the tornado track.
Furthermore, two of the tornadoes that have been rated EF5 featured no standard DIs meeting EF5 criteria. While an initial survey of the Joplin, Missouri, tornado of 22 May 2011 did identify 22 homes as having received EF5 damage (Marshall et al. 2012), an analysis performed by the American Society of Civil Engineers (ASCE) determined that none of those homes required EF5 wind speeds to account for the damage they received (Prevatt et al. 2012). However, Karstens et al. (2012) concluded that the EF5 rating was still justified based on the movement of rebarred parking stops along the track. The 24 May 2011 El Reno, Oklahoma, tornado was rated EF5 based on a combination of damage to the Cactus 117 oil rig in conjunction with data from the University of Oklahoma’s Rapid X-band Polarimetric Radar (RaXPol; Pazmany et al. 2013) but with no official EF5 DIs (Ortega et al. 2012).
The use of nonstandard DIs is not unique to EF5 tornadoes nor is it without precedent in the era of the legacy F scale. The EF4 rating in the Bowdle, South Dakota, tornado of 22 May 2010 was based on a nonstandard application of DI 24 (transmission line tower), when a high-tension transmission tower was tossed several hundred meters (NWS Aberdeen 2023). The collapse and removal of a high-tension transmission tower exceeds the maximum standard DOD for DI 24 on the EF scale (Karstens et al. 2012), which at most can yield an EF3 rating. Nevertheless, the damage to the transmission tower was rated EF4. The Plainfield, Illinois, tornado of 28 August 1990 was rated F5 by Fujita (1993) based solely on damage to a corn field northwest of town. The justification for rating the corn damage northwest of Plainfield as F5 was that it was significantly worse than corn damage found adjacent to structural damage rated F4. Another case of an F5 rating being applied through a nonstandard DI occurred on 6 May 1973 at Valley Mills, Texas. This tornado was rated F5 based on a report of a pickup truck being thrown approximately 800 m (0.5 mi; Grazulis 1993). No F5 structural damage occurred in either the Plainfield or Valley Mills tornadoes. NWS guidance still recommends the consideration of context and other evidence in determining a final rating, but outside of explicitly prohibiting the use of mobile radar data in assigning a final rating, the guidance does not directly address how to weigh the importance of nonstandard DIs (NWS Storm Data Directives 10-1605 2024).
Ultimately, the lack of any EF5-rated tornadoes since the 2013 Moore tornado represents a statistically anomalous gap in F/EF5 ratings over the past decade relative to the start of the record established by Grazulis (1993) dating back to 1880 (Fig. 2). This gap is the longest such gap in (E)F5 ratings in the historical record, despite there being at least four distinct periods of differing predominant rating methods: 1) the 1880–1949 period, where ratings were applied by Grazulis by collection of historical documentation; 2) the 1950–mid-1970s period, where students initially rated most of the tornadoes based on historical documentation, which was then edited by Grazulis and the National Weather Service; 3) the mid-1970s–2006, where F-scale ratings were applied by the NWS in real time; and 4) 2007–the present, during the EF-scale rating era (Edwards et al. 2013). During the entire 1880–2013 period, inclusive of the 2000–06 period with a decrease in violent tornado ratings noted by Edwards et al. (2021), 59 of the 134 calendar years featured at least one F5/EF5-rated tornadoes, yielding a calendar year probability of at least one F5/EF5 tornado occurring of 44.0%. Conversely, the probability of no F5/EF5 tornadoes in a given calendar year over the 1880–2013 period was 56.0%. Adding in the 2014–23 period and recalculating the probability of at least one F5/EF5 being observed in a calendar year for the 1880–2023 period reduces the probability from 44.0% to 41.0%. If the occurrence of at least one F5/EF5 tornado in a given calendar year is assumed to be independent from year to year, then the probability of 10 consecutive years not containing an F5/EF5 tornado in the United States, as observed across 2014–23, would be 0.3%.
Bar graph of F/EF5 tornado counts per year from 1880 to 2023.
Citation: Bulletin of the American Meteorological Society 106, 8; 10.1175/BAMS-D-24-0066.1
3. “EF5 candidates”
One possible way to evaluate if there were a change in what is considered an “EF5 tornado” would be to apply the previous standard for rating home damage as F5: ensure that the sweeping of a site-built home, built to code, off of its foundation constitutes EF5 damage. To evaluate how this adjustment would impact the climatology of EF5 ratings, tornadoes that have been rated EF4 but with final wind speed estimates (from Storm Data) of 84.9–89.4 m s−1 (190–200 mph) were identified and given the status of EF5 candidates (Table 2). A 4.5 m s−1 (10 mph) buffer below the 89.4 m s−1 (200 mph) expected value wind speed was applied to account for possible downward evaluation. These tornadoes were checked against information from the Damage Assessment Toolkit (DAT; NWS 2023) when DAT information was available and consistent with the official Storm Data narrative.
List of all tornadoes in the EF-scale era rated EF4 with final maximum wind gust estimates of 84.9–89.4 m s−1 (190–200 mph), the peak DI/DOD combination from Storm Data or the DAT, and the deviation of the applied maximum wind gust from the expected value wind gust of the peak DI/DOD.
To assess the climatological impact of what would happen if these tornadoes received EF5 ratings instead of their official EF4 ratings, Fig. 2 was reproduced by adding all the tornadoes from Table 2 to the official F/EF5-rated tornadoes from 1973 to 2023 (Fig. 3). Then, the probability of an F5/EF5 tornado being observed in a calendar year for the period 1880–2023 was recalculated with the EF5-candidate tornadoes treated as EF5s. If all EF5-candidate tornadoes were instead rated EF5, the calendar year probability of observing an F5/EF5 tornado from the 1880–2023 dataset would be 55.6%, nearly equivalent to the probability of an F5/EF5 tornado being observed in a calendar year as derived from the 1880–2013 dataset (56.0%) prior to the ongoing 10-yr EF5 gap.
As in Fig. 2, but for all F/EF5 tornadoes from 1973 to 2023 and all EF4-rated tornadoes with maximum estimated wind gusts of 84.9–89.4 m s−1 (190–200 mph) during the EF era.
Citation: Bulletin of the American Meteorological Society 106, 8; 10.1175/BAMS-D-24-0066.1
4. What should “EF5” represent?
While either lowering the threshold for EF5 from 89.9 m s−1 (201 mph) to 84.9 m s−1 (190 mph) or increasing the estimated wind speeds with all tornadoes in Table 2 to >200 mph to include the EF5 candidates as EF5s would create a more consistent F/EF5 climatology from 1973 to the present, it remains a valid question as to whether or not that is what a 5 rating “should” represent. According to the NWS Storm Prediction Center’s (SPC) One Tornado (ONETOR) database (Schaefer and Edwards 1999), 28 of the 17 375 tornadoes rated F1 or greater from 1 January 1973 (a threshold and analysis period used due to its relative long-term stability in tornado records; Coleman and Dixon 2014) through 31 January 2007 (the real-time rating period of legacy F-scale record) were rated F5 or 0.16%. From the implementation of the EF scale on 1 February 2007 through the end of calendar year 2022, that fraction dropped to 9/8626 or 0.10%. If the 2011 Joplin and El Reno tornadoes were to be downgraded in the dataset, given their lack of standard EF5 DIs, the fraction would drop to 7/8626 (0.08%) or half the 1973–early 2007 frequency. However, inclusion of Joplin and El Reno 2011 in addition to the EF5 candidates prior to 2023 would increase the frequency to 23 out of 8626 or 0.27%.
The singular reason that both the EF-era frequency of tornadoes rated EF5 is even as close as it is to the F-scale era—and that the inclusion of EF5 candidates would lead to a frequency of 5-rated tornadoes that would be much higher than 1973–early 2007—is the extreme nature of the 26–28 April 2011 super outbreak. If the 192 tornadoes exceeding EF1 intensity, including the four EF5-rated tornadoes and five EF5-candidate tornadoes of that outbreak (Lyza et al. 2022), were removed and these statistics were recalculated, that would leave 5/8434 EF1+ tornadoes (0.06%) from 1 February 2007–31 December 2022 being rated EF5. Removing 26–28 April 2011 and downgrading the Joplin and El Reno 2011 tornadoes would result in 3/8434 (0.04%) EF1+ tornadoes being recorded as EF5s, and retaining Joplin and El Reno 2011 while adding in EF5 candidates that did not occur on 26–28 April 2011 would yield 14/8434 (0.17%) EF1+ tornadoes receiving EF5 ratings or a frequency nearly identical to (0.01% greater than) the real-time F-scale era frequency of F5 tornadoes among the F1+ tornado population.
But ultimately, the meteorology, engineering, and social science communities need to ask a key question: what should EF5 represent? Is the real-time F-scale rating era, where F5 ratings were awarded to between 0.1% and 0.2% of all F1+ tornadoes, the best baseline for what a “5” rating should represent? Should an EF5 rating be tied directly to an 89.9 m s−1 (201 mph) wind speed threshold, or should it be lowered to better match the climatological use of single-family homes as F5 DIs, to retain the consistency in rating meaningfulness that was originally stated as an intention in the deployment of the EF scale (McCarthy et al. 2006)? What problems might such a change pose given the findings of Lyza et al. (2024), which indicate that the EF-scale wind speed ranges likely underestimate the near-surface wind speeds of strong–violent tornadoes, regardless of the density or type of DIs impacted? Is there a usefulness in having a rating based on wind speeds that very few structures can yield? Should there be a pathway for when damage observed beyond the maximum standard DOD of a standard EF-scale DI (e.g., Bowdle 2010) or to nonstandard DIs altogether (e.g., Plainfield 1990, Joplin 2011, and El Reno 2011) is permitted to be assigned an EF5 rating? Or, perhaps most radically, should tornado ratings be more reflective of total impact and not solely tied to wind speed estimates?
Acknowledgments.
We thank Jim LaDue for conversations that helped inform this work. Alan Gerard and three anonymous reviewers provided thoughtful reviews that greatly improved the quality of this manuscript. The statements, findings, conclusions, and recommendations are those of the authors and do not necessarily reflect the views of NOAA or the U.S. Department of Commerce.
Data availability statement.
NWS DAT data can be accessed online (https://apps.dat.noaa.gov/stormdamage/damageviewer/) and official Storm Data information can be accessed in a tabular format from the Storm Prediction Center’s ONETOR database (Schaefer and Edwards 1999; available online at https://www.spc.noaa.gov/wcm/).
References
Burgess, D., and Coauthors, 2014: 20 May 2013 Moore, Oklahoma, tornado: Damage survey and analysis. Wea. Forecasting, 29, 1229–1237, https://doi.org/10.1175/WAF-D-14-00039.1.
Coleman, T. A., and P. G. Dixon, 2014: An objective analysis of tornado risk in the United States. Wea. Forecasting, 29, 366–376, https://doi.org/10.1175/WAF-D-13-00057.1.
Edwards, R., J. G. LaDue, J. T. Ferree, K. Scharfenberg, C. Maier, and W. L. Coulbourne, 2013: Tornado intensity estimation: Past, present, and future. Bull. Amer. Meteor. Soc., 94, 641–653, https://doi.org/10.1175/BAMS-D-11-00006.1.
Edwards, R., H. E. Brooks, and H. Cohn, 2021: Changes in tornado climatology accompanying the enhanced Fujita scale. J. Appl. Meteor. Climatol., 60, 1465–1482, https://doi.org/10.1175/JAMC-D-21-0058.1.
Fujita, T. T., 1971: Proposed characterization of tornadoes and hurricanes by area and intensity. University of Chicago SMRP Research Paper 91, 42 pp.
Fujita, T. T., 1992: Memoirs of effort to unlock the mystery of severe storms during the 50 years, 1942-1992. Wind Research Laboratory, Research Paper Number 239, 304 pp.
Fujita, T. T., 1993: Plainfield tornado of August 28, 1990. The Tornado: Its Structure, Dynamics, Prediction, and Hazards, Geophys. Monogr., Vol. 79, Amer. Geophys. Union, 1–17, https://doi.org/10.1029/GM079p0001.
Grazulis, T. P., 1993: Significant Tornadoes, 1680–1991: A Chronology and Analysis of Events. Environmental Films, 1326 pp.
Karstens, C. D., W. A. Gallus Jr., P. Sarkar, and T. P. Marshall, 2012: Supplemental damage indicators discovered in recent strong tornadoes. 26th Conf. on Severe Local Storms, Nashville, TN, Amer. Meteor. Soc., 106, https://ams.confex.com/ams/26SLS/webprogram/Paper211981.html.
Lyza, A. W., M. D. Flournoy, and E. N. Rasmussen, 2022: Observed characteristics of the tornadic supercells of 27–28 April 2011 in the southeast United States. Mon. Wea. Rev., 150, 2883–2910, https://doi.org/10.1175/MWR-D-21-0274.1.
Lyza, A. W., M. D. Flournoy, and A. A. Alford, 2024: Comparison of tornado damage characteristics to low-altitude WSR-88D observations and implications for tornado intensity estimation. Mon. Wea. Rev., 152, 1689–1710, https://doi.org/10.1175/MWR-D-23-0242.1.
Marshall, T. P., 2002: Tornado damage survey at Moore, Oklahoma. Wea. Forecasting, 17, 582–598, https://doi.org/10.1175/1520-0434(2002)017<0582:TDSAMO>2.0.CO;2.
Marshall, T. P., K. Jungbluth, and A. Baca, 2008a: The Parkersburg, IA tornado. 24th Conf. on Severe Local Storms, Denver, CO, Amer. Meteor. Soc., P3.3, https://ams.confex.com/ams/24SLS/webprogram/Paper141533.html.
Marshall, T. P., D. McCarthy, J. G. LaDue, J. Wurman, C. R. Alexander, P. Robinson, and K. A. Kosiba, 2008b: Damage survey and deduction of vortex structure of the Greensburg, KS tornado. 24th Conf. on Severe Local Storms, Denver, CO, Amer. Meteor. Soc., 8B.3, https://ams.confex.com/ams/24SLS/techprogram/paper_141534.htm.
Marshall, T. P., W. Davis, and S. Runnels, 2012: Damage survey of the Joplin tornado. 26th Conf. on Severe Local Storms, Nashville, TN, Amer. Meteor. Soc., 6.1, https://ams.confex.com/ams/26SLS/webprogram/Paper211662.html.
McCarthy, D. W., J. T. Schaefer, and R. Edwards, 2006: What are we doing with (or to) the F-scale? 26th Conf. on Severe Local Storms, Kansas City, MO, Amer. Meteor. Soc., 5.6, https://ams.confex.com/ams/23SLS/techprogram/paper_115260.htm.
Mehta, K. C, and R. Carter, 1999: Assessment of tornado wind speed from damage in Jefferson County, Alabama. Wind Engineering into the 21st Century: Proceedings of the Tenth International Conference on Wind Engineering, CRC Press, 265–271.
Minor, J. E., J. R. McDonald, and K. C. Mehta, 1977: The tornado: An engineering oriented perspective. NOAA Tech. Memo. ERL NSSL-82, 103 pp.
NCEI, 2023: Storm events database. NCEI, accessed 5 November 2023, https://www.ncdc.noaa.gov/stormevents/.
NWS, 2023: Damage assessment toolkit. NWS, accessed 7 November 2023, https://apps.dat.noaa.gov/StormDamage/DamageViewer/.
NWS Aberdeen, 2023: May 22, 2010 South Dakota tornadoes. NWS, accessed 5 November 2023, https://noaa.maps.arcgis.com/apps/MapJournal/index.html?appid=494a6e79428c46cc9e9405a87df48373.
NWS Birmingham, 2023: Hackleburg (Marion County) EF5 tornado: April 27, 2011. NWS, accessed 6 November 2023, https://www.weather.gov/bmx/event_04272011hackleburg.
NWS Jackson MS, 2024: NWS Jackson, MS April 25-27, 2011 Severe Weather Outbreak—Neshoba/Kemper/Winston/Noxubee Counties EF5 tornado. NWS, accessed 11 June 2024, https://www.weather.gov/jan/2011_04_25_27_neshoba_kemper_winston_noxubee.
NWS Storm Data Directives 10-1605, 2024: Storm data preparation. NWS, 110 pp., https://www.weather.gov/media/directives/010_pdfs/pd01016005curr.pdf.
Ortega, K. L., and Coauthors, 2012: Overview of the 24 May 2011 tornado outbreak. Special Symp. on the Tornado Disasters of 2011, New Orleans, LA, Amer. Meteor. Soc., J4.2, https://ams.confex.com/ams/92Annual/webprogram/Paper201297.html.
Pazmany, A. L., J. B. Mead, H. B. Bluestein, J. C. Snyder, and J. B. Houser, 2013: A mobile Rapid-Scanning X-band polarimetric (RaXPol) Doppler radar system. J. Atmos. Oceanic Technol., 30, 1398–1413, https://doi.org/10.1175/JTECH-D-12-00166.1.
Phan, L., and E. Simiu, 1998: The Fujita tornado intensity scale: A critique based on observations of the Jarrell tornado of May 27, 1997. NIST Tech. Note 1426, 20 pp., NTIS PB98-175136INZ.
Prevatt, D., W. Coulbourne, A. Graettinger, S. Pei, R. Gupta, and D. Grau, 2012: Joplin, Missouri, Tornado of May 22, 2011: Structural Damage Survey and Case for Tornado-Resilient Building Codes. American Society of Civil Engineers, 54 pp., https://doi.org/10.1061/9780784412503.
Schaefer, J. T, and R. Edwards, 1999: The SPC tornado/severe thunderstorm database. 11th Conf. on Applied Climatology, Dallas, TX, Amer. Meteor. Soc., 603–606.
U.S. Department of Commerce, 2002: Service assessment: La Plata, Maryland, tornado outbreak: April 28, 2002. Accessed 5 November 2023, https://www.weather.gov/publications/assessments.
WSEC, 2006: A recommendation for an Enhanced Fujita scale (EF-scale). WSEC Texas Tech University, 111 pp., https://www.depts.ttu.edu/nwi/Pubs/EnhancedFujitaScale/EFScale.pdf.
