Tornado Deaths: Limits of Current Technology

An earlier draft of this article first published as Tornado Deaths and the Limits of Tornado-Related Technologies onTechnorati.

Deaths in the southeastern United States related to the April 27, 2011 tornadoes prompt an examination of tornado safety.  A key question is whether modern tornado detection and warning technology significant reduces risk of tornado-related death.

Storm and tornado tracking technology has exploded over the last 30 years.  Doppler radar technology continues to improve and national and regional news channels trumpet their own systems with names such as SkyTraK DopplerMax, First Alert Mega Doppler Radar and Doppler 9000 HD.  Local news stations in the the midwestern  and southern U.S. regions frequently break into regular programming to alert viewers to potential dangerous storms and tornado watches.

In 1982, The Weather Channel started broadcasting in the United States.  This 24 hour channel monitors severe weather across the U.S. providing access to instant weather information.

But what has been the effect of these tornado technologies in the trends for tornado-related deaths?  The trends to do not appear to support these technologies in their ability to reduce tornado-related death.  Below is a plot of the yearly number of tornado-related deaths in the U.S. between 1950 and 2011 (year-to-date) from data published by the U.S. National Oceanic and Atmospheric Administration. 

Tornado-related deaths have dropped dramatically in the United States but the biggest declines occurred between 1925 and 1980.  Between 1875 and 1925, the U.S. averaged nearly two deaths per million residents.  Since 1980, the U.S. death rate has averaged around .2 deaths per million people.  This is around a 90 per cent decrease and a significant advance in public health safety.

The problem for tornado technology advocates is the trend in tornado-related deaths between 1980 and 2011.  The over 300 deaths in the recent southeastern U.S. tornado make 2011 already one of the top three years for tornado deaths in the U.S. since 1950.  The number of deaths so far in 2011 is topped by only by 1953 with 519 deaths and 1974 with 366 tornado-related deaths .  Yearly death rates vary significantly and should be corrected for total population to compare trends.  Taking a look at the rates of tornado-related deaths per million population finds the following results in the last six decades:

Decade      Total Deaths           Deaths/Million

1950s        1419                                  8.6

1960s          942                                 4.9

1970s          998                                 4.7

1980s          522                                 2.2

1990s          579                                 2.2

2000s          556                                 1.9
2010s          420

The decline in the 2000s to 1.9 deaths/million/decade is not statistically different than the rate for the decade of the 1980s or the rate for the 1990s decades.  The 300 deaths in the southeastern U.S. in the recent U.S. tornado is about the number that could be expected over a six year span in the U.S.  The likelihood that the tornado deaths in the 2010s will show a statistically significant decline seem small.


There is some data supporting regional tornado death decreases in the area surrounding new radar installations  Additionally, there is some support for increased numbers of residents and mobile homes in the southeastern U.S. potentially contributing to the plateau in number of tornado deaths in the U.S. Nevertheless, it's hard to look at the raw data and feel like tornado detection technology has dramatically reduced the risk of tornado death.

Technology has contributed to improvement in many areas of life.  Tornado-related deaths in the U.S. have dropped significantly in the U.S.  The timing of this decline suggests factors such as improved building safety may have contributed to this public health improvement.

However, the data do no support modern tornado detection and warning technology as  major factors in decreasing tornado-related deaths.   Without significant new technological advances, it appears a limit has been reached in reducing tornado-related deaths in the U.S.


The figure in this post is created by the author from public domain data.

Ashley, W. (2007). Spatial and Temporal Analysis of Tornado Fatalities in the United States: 1880–2005 Weather and Forecasting, 22 (6), 1214-1228 DOI: 10.1175/2007WAF2007004.1

Simmons, K., & Sutter, D. (2005). WSR-88D Radar, Tornado Warnings, and Tornado Casualties Weather and Forecasting, 20 (3), 301-310 DOI: 10.1175/WAF857.1


Brooks, Harold E (2002). Deaths in the 3 May 1999 Oklahoma City Tornado from a Historical Perspective Weather and Forecasting


The Science of Tornadoes--3 minute video from the History Channel

Suicide Tops Death List for Meth Heads

Methamphetamine Chemical Structure

Methamphetamine dependence can lead to an early death.  The magnitude and mechanism of this effect is not well understood.  One way to better understand effects of drug abuse/dependence on mortality is the prospective outcome study.  These types of studies tend to be costly and may require many years to yield research results.   That’s why there are not many published studies to answer the question of this post.

A recent study from Taiwan provides some valuable insight into this issue.  This study followed a cohort of 1254 individuals with a history of methamphetamine abuse and a psychiatric admission for treatment between 1990 and 2007.  National death records were queried and the methamphetamine abuse cases were compared to an age- and gender- matched control group.  One hundred thirty methamphetamine users died during follow up with the following leading categories of death:  

• Suicide n=42 (32.3%)
• Accidents n=26 (20.0%)
• Undetermined unnatural deaths n=14 (10.8%)
• Cardiovascular disease n=13 (10.0%)
• Undetermined natural deaths n=9 (6.9%)
• Liver disease n=6 (4.6%)

In this Taiwanese sample, methamphetamine abuse carried an overall six fold increase in mortality.  Unnatural deaths (suicides, accidents, homicides, undetermined natural deaths) were particularly elevated.  Male amphetamine abusers had about a 10-fold increase in unnatural deaths while female methamphetamine abusers had a remarkable 26-fold increase in unnatural death during follow up.

Being married appeared to reduce the risk of death from unnatural and natural causes in this cohort.  Use of other substances in addition to methamphetamine increased risk of death.

I would suspect that in the United States, homicide would be a greater contributor to mortality in those with a history of methamphetamine dependence.  Overall homicide rates in Taiwan are much lower than in the United States and may reduce risk for this type of unnatural death.

Interestingly, about two thirds of the cohort had exhibited methamphetamine psychosis-a psychiatric complication of methamphetamine use.  Those with a history of this type of psychosis were no more likely to die than those without a history of the psychosis.

One problem with this type of study is difficulty assessing the specific effect of methamphetamine from other potential confounding factors.  For example, methamphetamine use is higher in smokers than smokers and so some increased mortality risk could be related to effects of smoking.  Methamphetamine dependence is more prevalent in a variety of mental disorders linked to increased suicide.

Nevertheless, this study shows the magnitude and type of mortality risk linked to methamphetamine dependence.  Public health interventions that reduce the prevalence of methamphetamine dependence would like reduce the excess mortality associated with this drug.

Wikipedia Commons chemical structure of methamphetamine image authored by Harbin.

Kuo CJ, Liao YT, Chen WJ, Tsai SY, Lin SK, & Chen CC (2010). Causes of death of patients with methamphetamine dependence: A record-linkage study. Drug and alcohol review PMID: 21355920