by: Thomas W. Schmidlin, Department of Geography, Kent State University
A tornado is a deadly phenomenon that strikes with little warning and can destroy a building in a matter of seconds. Thus, knowledge of the attributes of persons killed by tornadoes, their behavior when the storm threatened, and the circumstances of their death are useful in evaluating hazard preparedness, safety rules, and warning methods. This information identifies high-risk groups, high-risk situations, and high-risk behaviors and can be used to improve preparedness and warning programs and reduce tornado-related deaths.
Our research during the past five years has focused on risk factors for death due to tornadoes in the United States. My colleagues, Paul King, Barbara Hammer, and Yuichi Ono, and I have asked the question, "Why do some people die in tornadoes while others survive?" To find answers, we travel to the site of a tornado disaster about one week after the event in order to complete a detailed study of each fatality and of survivors who were in the path of the tornado. This research has been funded by Quick Response grants made available by the Natural Hazards Research and Applications Information Center (see the Observer, Vol. XXII, No. 1, p. 5).
Our survey quantifies demographic data and information related to method of warning, access to warning, time of awareness of the impending tornado, exact location when the storm struck, degree of destruction at the site, and so on. Survivors are interviewed in person. Surveys are completed for fatalities through interviews with relatives, supplemented by neighbors, coroners, and funeral home directors. The survey responses for those who died are then compared to responses from those who survived, in order to identify differences between the two groups.
Data have been collected for the Georgia and Alabama tornadoes of March 27, 1994, and the Arkansas tornadoes of March 1, 1997. Surveys were completed for 45 fatalities and 104 survivors. In both cases, several tornadoes struck across rural areas of a southern state on a weekend afternoon with 10 to 30 minutes warning time from the National Weather Service (NWS).
What We Learned
Results from the Georgia and Alabama tornadoes (Disasters 19 (1995): 170-177) showed risk factors for death to include advanced age, location in a mobile home, location in a room above ground with windows, not watching television in the hour before the tornado hit, and being aware of the approaching tornado for less than one minute. Results from the Arkansas tornadoes (available from the Natural Hazards Center as Quick Response Report #98, see the article in this Observer) also showed risk factors for death to be location in a mobile home and in a room above ground with windows. In contrast to the earlier study, there was no difference in age between fatalities and survivors, although being divorced appeared as a risk factor, possibly due to the isolation and reduced income of divorced persons.
These results generally reinforce previous assumptions that were developed from studies of tornadoes and other hazards:
The supreme importance of protection by a building in preventing deaths was evident. Mobile homes and outer rooms of frame homes do not offer that protection. In spite of long warning lead times, most people--fatalities as well as survivors--first became aware of the approaching tornado only when they saw or heard it, allowing little time to reach shelter. In addition, television was used to obtain weather information much more than radio. None of the persons we interviewed used NOAA weather radio on the day of the tornadoes.
Additional research of this type will provide a composite of tornado risk factors over a variety of geographic, demographic, and cultural settings--the foundation of a stable and reliable database from which general conclusions may be drawn.
Are Cars Safer than Mobile Homes?
While conducting the first study in Georgia and Alabama in 1994, we were surprised by the common occurrence of cars or pickup trucks that remained upright with little damage near mobile homes that were destroyed and the mobile home occupants killed. After careful thought, it did not seem so surprising. After all, a modern car has a low center of gravity, a streamlined form, a protective interior, and is designed to encounter strong winds and protect occupants in case of a roll-over and other crashes. Our preliminary estimates showed that a door-handle-height wind speed of about 120 mph is required to tip a car, compared to perhaps 80 mph to tip a mobile home.
Rural mobile home residents have few options when a tornado threatens. Underground shelters are rare and a sturdy building for shelter is usually some miles away. In those desperate situations when sturdy shelter is not within running distance from a mobile home, both the NWS and the American Red Cross recommend that mobile home residents leave the mobile home and "lie flat in a ditch or low-lying area" when a tornado warning is issued.
Mobile home residents recognize the legendary vulnerability of their dwellings in wind storms. However, when the tornado siren starts blowing or the Weather Channel screen turns red, few are willing to gather the family and leave their mobile home to run outside into a severe thunderstorm with heavy rain, lightning, hail, and flying debris to lie down in a water-filled ditch to await a tornado. They tell us such actions are counterintuitive.
Following our field observations in 1994, a reasonable option for those in mobile homes without nearby shelter seemed to be to drive to one. Our public statements in 1994 that rural mobile home residents with no nearby shelter may be safer getting into their vehicles and driving to a shelter when a tornado threatens, rather than running outside to lie down in the storm, drew widespread media attention, many comments of agreement, and strong comments to the contrary by a few people in the NWS.
In light of NWS and Red Cross recommendations that mobile home residents and vehicle occupants exit and lie down outdoors when a tornado threatens, we sought previous studies on the relative safety of being in a vehicle compared to being outdoors that supported those recommendations. As we reported in a commentary last year (Bulletin of the American Meteorological Society 77: 963-964), no studies have been found to support those recommendations.
Tornado Strength and Vehicle Safety
Following our surprising observations after the 1994 tornadoes, we embarked on a systematic survey of the effects of tornadoes on cars and pick-ups. We collected data on vehicles parked outdoors at homes with F1, F2, or F3 tornado damage following the Louisville tornado in May 1996, the Arkansas tornadoes in March 1997, and the Texas tornadoes in May 1997.
Not surprisingly, this sample of 180 vehicles showed that the percentage of cars moved or tipped tended to increase with increased home damage (and inferred wind speed). Surprisingly, at homes with F3 damage (158- 206 mph), fewer than half (46%) of cars were moved by the wind, only 15% were tipped over by the wind, and 39% of the vehicles were damaged sufficiently to cause serious injury to potential occupants. These results are now under review for publication.
Where does that leave us with respect to surviving tornadoes? We will continue postdisaster research of deadly tornadoes to determine general principles of high-risk behaviors and identify high-risk groups. It is clear that the 73 mph or less wind speeds of F0 tornadoes, the weakest of weak tornadoes, pose little threat to human life. It is also clear that the rare violent tornadoes (F4 and F5) with maximum wind speeds over 206 mph will destroy well-built homes and toss vehicles. The only reasonable protection in these extreme cases is an underground shelter, but only 3% of tornadoes have these wind speeds.
It is in the middle range of F1, F2, and F3 wind speeds that most tornadoes occur. Underground shelter is always safest, but the interior rooms of well-built homes or offices provide life-saving shelter in most cases. At the same time, mobile homes clearly remain a high-risk location in this range. For nearly half of the Americans who die from tornadoes, the last view they have of this world is the disintegrating interior of their mobile home. Only one-third of the 15 million mobile home residents in the U.S. live in a mobile home park, and some of these do not have sturdy shelters for all residents. The other 10 million live on private rural land,
and many of these people will not have a sturdy shelter within running distance when the tornado warning is issued. Mobile home occupancy is predicted to increase for the foreseeable future, and millions of Americans are on the road in their cars and trucks during the late afternoon when tornadoes are most likely to occur.
Conclusions
The hazards community has an opportunity to find reasonable, affordable, and practical means of reducing the risk of death to mobile home residents due to tornadoes. We must also strive to provide safety recommendations for mobile home residents and vehicle occupants that are based on modern research.
The author can be contacted at Kent State University, P.O. Box 5190, Kent, OH 44242-0001; (330) 672-2045; fax: (330) 672-4304; e-mail: tschmidl@kent.edu.
Return to the SFTS! Severe Weather Safety and Awareness Page