Edited Version November 11, 1998 Transcript

EIIP Virtual Library Online Presentation

"Risk Factors for Death in the February '98 Florida Tornadoes"

Thomas Schmidlin
Professor and Chair, Department of Geography
Kent State University

EIIP Moderator: Amy Sebring

The original unedited transcript of the November 11, 1998 online Virtual Library presentation is available in EIIP Virtual Forum Archives (http://www.emforum.org). The following version of the transcript has been edited for easier reading and comprehension. Typos were corrected, date/time/names attributed by the software to each input were deleted but the content of questions and responses are as stated by each participant. Answers from the participants to questions by the audience are grouped beneath the appropriate question to facilitate meaning.


Amy Sebring: Welcome to the EIIP Virtual Library! We are pleased to have Dr. Thomas Schmidlin, Kent University, rescheduled for a Virtual Library presentation. Last month, we had a little problem, but Dr. Schmidlin graciously agreed to come back today. ...


Thomas Schmidlin is Professor and Chair of the Department of Geography at Kent State University. Tom received a B.S. degree from Iowa State University, M.S. from the University of Nevada-Reno, and the Ph.D. in Atmospheric Sciences from Cornell University. His research focuses on tornado hazards.

Tom and his colleagues have conducted fieldwork following tornado disasters throughout the United States and have presented their research at numerous scientific meetings in the U.S. and in Europe and to state and county EMA. His research on tornado hazards has been published in Disasters, Bulletin of the American Meteorological Society, and by the American Geophysical Union and the Natural Hazards Center.

Tom is Editor of The Ohio Journal of Science and he and his wife Jeanne are co-authors of the book, Thunder in the Heartland: A Chronicle of Outstanding Weather Events in Ohio. The paper Tom will present is titled "Risk Factors for Death in February 98 Florida Tornadoes."

After the formal presentation, we will have time for Q&A. We will give you instructions for that part of today's session after the presentation. And now, Dr. Thomas Schmidlin --- thank you, Tom, for joining us today. We welcome you and turn the floor to you now.


Tom Schmidlin: Thank you Amy. We are holding our own in the winds of Ohio today!

Risk Factors for Death in the 22-23 February 1998 Florida Tornadoes, Thomas W. Schmidlin, Paul S. King, Barbara O. Hammer, Yuichi Ono.

Tornadoes on the night of 22-23 February 1998 killed 42 people in central Florida. We conducted field research in the disaster areas to assess risk factors associated with death among persons in the tornado paths. If we know some information about the persons killed by tornadoes, their behavior as the tornado approached, and the circumstances of death, we can compare that to information about those who were not killed to evaluate tornado preparedness programs, safety rules, and warning methods. The goal is to reduce tornado injuries and deaths.

Regional differences in death rates might be caused by differences in tornado severity, urbanization, building construction, preparedness, hospital facilities, warning systems, and the distinctive behavioral characteristics of individuals. To get detailed information on those who die in tornadoes and those who survive, we started research several years ago to collect information soon after tornado disasters.

In the Georgia Tornadoes of March 1994 we found 15 of the 20 deaths were in rural mobile homes. Those who died were older than survivors, more likely to be in a room above ground with windows, less likely to be listening to the television, had less warning time, and more likely to have been struck by an object than survivors. Although NWS warnings preceded all of the Georgia killer tornadoes and 2/3 of survivors had a radio or TV on in the hour before the tornadoes struck, 68% became aware of the approaching tornado only by seeing or hearing the tornado. This gave little time for action.

Our research on the March 1997 tornadoes in Arkansas showed 14 of the 26 deaths were in mobile homes. Risk factors for death were being in a room above ground with windows and being in a room that lost walls, roof, or floor. About 61% of the Arkansas survivors first became aware of the tornado when they saw or heard the tornado, giving little time for protective action. We found vehicles parked outside damaged homes to be remarkably stable, even with extensive house damage.

The purpose of this spring's Florida research was to obtain information on persons who died in the February tornadoes and those who survived, to build upon our 1994 field work in Georgia and our 1997 field work in Arkansas. At least seven tornadoes occurred across central Florida during the night of 22-23 February 1998. There were 42 deaths and about 260 persons injured.

Tornado watches had been issued by the NWS Storm Prediction Center for central Florida all afternoon and evening on 22 February. The watches mentioned that conditions were favorable for the development of supercell thunderstorms. The first deadly tornado (F2) hit in Volusia County at 22:55 (10:55 PM) and killed one person and injured three along an 11 km path. The National Weather Service (NWS) in Melbourne issued a tornado warning for Volusia County at 22:41.

Another tornado (F3) along a 32 km path in Lake and Orange counties from 23:37 to 00:00 killed three and injured 70 people. The NWS issued tornado warnings for Lake County at 23:27, for western Orange County at 23:37, and then Seminole County at 23:46. A second tornado touched down at 00:10 (12:10 AM) in Seminole County, killing 13 and injured 36 others along the 22 km path. Seminole County was under a tornado warning for this storm from 23:46 until 00:55. A tornado in Osceola County at 00:40 traveled 56 km, killing 25 in Osceola County and injuring about 150 persons. A tornado warning was issued for Osceola County at 00:22 effective until 01:20.

All counties where fatalities occurred were under tornado warnings at the time the tornadoes struck, giving warning lead times of 10 to 24 minutes. There were no community tornado warning sirens in the areas struck by the tornadoes.

We conducted four days of field work with Paul King from Cornell University and Kent State University Geography graduate students Barbara Hammer, Yuichi Ono and Heather Heckman. The research was funded by the Natural Hazards Center. A survey was completed for each fatality by obtaining information from medical examiners, funeral home directors, relatives, or neighbors. A survey was also completed for 86 survivors who were in the same structure with someone who died or at sites adjacent to the fatalities.

Damage at the sites of deaths was evaluated with the Fujita Tornado Intensity Scale. We used statistical techniques to test whether the responses to the survey questions differed between those who were killed and those who survived. Our results showed there were no deaths in frame houses during the Florida tornadoes, in spite of widespread F3 damage to occupied frame homes. Forty-one of the deaths (98%) were in mobile homes or parked recreational vehicles and one death (2%) was in a motor vehicle.

There was a significant difference in average age between fatalities (48 years) and survivors (39 years). There was no difference in gender between fatalities and survivors. Thirty-nine (93%) of the fatalities were white, two (5%) Hispanic, and one (2%) was black. In the survivor sample, 45 (52%) were white, 36 (42%) Hispanic, and five (6%) were black. There were significantly more whites among the fatalities than among the survivors. Among persons, 18 years and older, 56% of fatalities were currently married compared to 92% of the survivors. This is a significant difference.

No basements or underground storm shelters were seen or known in the region. All of the fatalities in buildings or RVs were in a room above ground with windows. In rooms where deaths occurred, 100% lost the ceiling, 100% lost a wall, 100% had the floor blown away, and 100% of deaths were hit by debris.

In rooms where survivors were located, 60% lost the ceiling, 44% lost a wall, 27% had the floor blown away, and 51% of survivors were hit by debris. Among the survivors, 47% were taking shelter behind or under furniture or a stairway and 39% used a rug or blanket to cover themselves when the tornado struck.

Most (68%) of the survivors first became aware of the tornado when they heard or saw the tornado, 10% first became aware of the tornado through a warning on the television, and 18% were told of the tornado by a friend, neighbor, or relative (typically a family member in the same house). Warning sirens did not exist in any of the communities we surveyed.

No one reported use of radio or NOAA weather radio to first receive the warning, although 32% of survivors reported a weather radio in their home. Only 2% of the survivors had a radio on during the hour prior to the tornado, while 34% reported watching the television in the hour prior to the tornado. Although tornado warnings had been issued 10 to 24 minutes prior to the first touchdown of each fatal tornado, 88% of survivors reported less than one minute awareness before the tornado struck. The remainder reported one to five minutes awareness.

These results clearly indicate the importance of protection by a building in preventing deaths. All deaths to persons in buildings occurred when above ground rooms disintegrated or collapsed in mobile homes and parked recreational vehicles.

Few of the survivors took the recommended action of hiding under heavy furniture or covering with a blanket. We also found that few survivors of the 1994 Georgia tornadoes and 1997 Arkansas tornadoes took these precautions, leading to the conclusion that these simple steps should be reinforced in tornado preparedness programs.

Government warnings preceded the tornadoes by 10 to 24 minutes but the occurrence of these tornadoes near midnight on a Sunday night reduced the effectiveness of radio and television in conveying tornado warnings to the public. Without community warning sirens or NOAA weather radio tone alerts, the only warning most people had was the noise of the approaching tornado and associated debris hitting their home.

Several large communities of mobile homes and RVs were struck by the tornadoes, leading to large numbers of deaths and injuries. Residents of these mobile home communities did not have nearby underground shelter. The local ‘shelter' in many peoples' minds was the community activity building. However, the sheltering value of these buildings is doubtful as they have large-span roofs, large glass windows, and did not appear to have the capacity to hold all the residents of the area.

Elderly residents told us their ‘tornado shelter' was the local school, several blocks away, but they stayed in their mobile homes because there was not time to reach the school and it would have been locked at midnight on a Sunday. They had apparently been instructed that the school was their ‘hurricane shelter' and assumed it would also serve as a tornado shelter. Actions to be taken during hurricanes and tornadoes need better distinction among Florida residents, especially the elderly.

With one exception, mobile home residents we surveyed did not leave their home to lie in a ditch or depression when they heard the tornado approaching. When we inquired why they did not take this step recommended by the National Weather Service and Red Cross, they responded that it was dark, they were in their night clothes, trees and wires were falling, things were blowing around, there were no ditches, ditches had alligators in them, and so on.

Our earlier research led us to propose that mobile home residents without nearby sturdy shelter may be safer in the car than going outside to lie down in a ditch or depression during a tornadic thunderstorm. Research continues on this topic.

The finding that persons of Hispanic ethnicity were at less risk of death than whites may be an artifact of the tornado path. The mobile homes struck by the tornadoes were largely occupied by whites, while there was a larger Hispanic population in the modern frame homes along the path.

This research found several results similar to those we reported earlier from Georgia and Arkansas tornadoes and a few contrasts with those results. Differences may result from the unique setting of each disaster.

We plan to continue collection of this ‘quick response' data to provide a stable and reliable data-base from a variety of tornado disasters. From this larger project, general conclusions may be drawn about tornado safety. Thank you for your interest and I look forward to some comments.

Amy Sebring: Thank you Tom. If you have a question for our speaker, please input a question mark (?) and wait for recognition from the moderator before sending your question. You may compose your question and have it ready to submit but do not send it until you are recognized.

[Audience Questions]


Isabel McCurdy: What conclusions can be drawn about tornado safety?

Tom Schmidlin: The strongest seem to be value of the structure in protecting from wind and debris and of the value of warning to provide time to seem shelter, and availability of shelter.


AK Miller: Did the previous findings on the relationships between mobile homes and autos hold true in Florida?

Tom Schmidlin: Yes, strikingly so - most of the mobile homes with deaths had intact vehicles.


Amy Sebring: I understand that they sold a lot of NOAA tone alert radios in Florida, AFTER the storms.

Tom Schmidlin: That was the only way for a person without TV on at midnight to get a warning.

Amy Sebring: I think your research highlights the need to do a better job on dissemination.

Tom Schmidlin: It is frustrating for the NWS, with good warning lead times but not effective in letting people know.


Isabel McCurdy: Was there any reason given why the survivors did not hide under furniture or cover with a blanket?

Tom Schmidlin: Nothing specific Isabel, maybe not enough time, or did not know the 'rules' on that.


Amy Sebring: You noted an average age difference. Is this similar to your previous findings, and was there much variation about the mean for those who died?

Tom Schmidlin: Yes Amy, we generally find that older person have a higher death rate. That is common to most other studies on tornado mortality. Causes may be frail health, reluctance to seek shelter. We have also seen that the very young, under one year, have a high death rate.


Amy Sebring: Did I also understand you to say there may also be a marriage factor?

Tom Schmidlin: Yes, this appeared for the first time in the Arkansas study last year. Divorced or 'not married' had a higher death rate. I do not have a clear reason for that. It could be demographics, more divorced people in mobile homes? I really do not know and would like suggestions.


Bill Murray: The underlying common factor in all results seems to be personal motivation.

Tom Schmidlin: Could you elaborate Bill?

Bill Murray: The recent IDNDR discussion on Prevention if read in one sitting shows the common thread is the failure of individuals to act on warnings UNLESS they personally related the danger to themselves. Married people agreed on danger; older people waiting for the warning to be validated by past experience; even those with NOAA radios waited. In other words, the underlying problem is MOTIVATING those who hear the warning.

Tom Schmidlin: Yes, Bill, there is some truth to that but many mobile home residents, there are 15 million in US, do not have an alternative shelter, underground or sturdy building nearby to go in the moments before the tornado strikes, and they are reluctant to go outdoors in a tornadic thunderstorm as NWS and Red Cross recommend, which is why they may stay in their mobile home.


Amy Sebring: Did you get to participate in the "Tornado Summit" Tom?

Tom Schmidlin: No, I wish I had the chance.


Isabel McCurdy: Are a lot of these people that live in these mobile homes, snow birds?

Tom Schmidlin: Most of those who were killed in mobile homes in Florida were not, they were permanent residents, although there were several fatalities among snowbirds in parked RVs in the Florida tornadoes, who also had no sturdy building available for shelter.


Amy Sebring: Tom do you make any attempt to compare events per capita? Total number of deaths in Arkansas to total in Florida, e.g., compared to the total of people at risk?

Tom Schmidlin: Well, Florida has the highest frequency of tornadoes of any state, per square mile, but most in the past have been weak, while Ark has a history of deadly violent tornadoes, and we might expect people in Ark to therefore, be more aware of tornadoes. We found some truth to that.


Amy Sebring: Did you receive any anecdotal reports as to why watches were ignored if they were?

Tom Schmidlin: People were aware of the watches that had been up all day, but were not aware of the warnings that were issued as the tornadoes touched down after many people went to bed.


Isabel McCurdy: Why are mobile homes allowed in an area when it is known that their ability to withstand tornadoes are poor?

Tom Schmidlin: It is matter of economics and risk. Tornadoes occur over much of the eastern 2/3 of US and to prohibit mobile homes is unrealistic, and to manufacture them to withstand tornadoes is uneconomical. The best is to have a personal shelter for mobile home residents, and this is why we have suggested a car is safer than a mobile home during a tornado, though NOT as safe as a sturdy building.


Bill Murray: Penalty for age is 30 years of experience. While there is no magic bullet, some legislative action may be effective. If all mobile home parks/RV parks., etc were required to provide shelter for residents it might contribute to saving lives while minimizing individual costs.

Tom Schmidlin: Absolutely Bill. The shelter needs to be a real shelter though. We found mobile home parks in Florida called the rec building or pool building the tornado shelter but were full of glass and no good at all.

Bill Murray: Thus we are back to motivation. It took many years for flood resistant standards and earthquake standards are still not fully implemented. Yet there is motivation in the political body for these. Perhaps we must devote more attention to convincing the leadership of the need

Tom Schmidlin: Plus over half of mobile homes in US are not in parks, but are rural stand-alone mobile homes, and in fact that is where most of the Georgia and Arkansas deaths occurred, rural mobile homes.

Bill Murray: Even in Montana the independent minded rural residents have reluctantly accepted imposition of flood plain restrictions. No structure can be erected, no matter how remote, without a flood plain certification.


Amy Sebring: Many thanks to our guest today, excellent job Tom!

Tom Schmidlin: Thanks Amy and I appreciated all the questions and comments.

Amy Sebring: Let me mention just a couple of upcoming items: tomorrow evening we will have our Mutual Aid session at 8:00 PM Eastern. And next Wed., we will have Merrily Powell from Essential Technologies in Tech Arena to talk about a new product.

Thank you all for coming today on a holiday, and the transcript will be posted next week.

Tom Schmidlin: Thanks Amy and all.