Lightning deaths by year (1940-2011) and tornado deaths by year (1950-2011) (Data from National Weather Service)

By comparison, 551 people were killed by tornadoes in 2011, the most since 1950 and third highest on record according to NOAA.

The striking difference between lightning and tornado fatalities begs the question why...

Why so few lightning deaths?

The relatively low number of lightning deaths resulted mainly from a successful public education campaign according to John Jensenius, the National Weather Service’s lightning expert.

“In large part, the reduction in lightning fatalities is due to the efforts of many people and organizations who help educate the public on the dangers of lightning and make the public more aware of our lightning safety recommendations,” he said.

(Author’s note: surely improved weather forecasts and warnings have also played a significant role)

Since 1940, lightning has killed 128 people per year on average. But the average dropped to just 37 per year over the last 10. By comparison, 323 people per year died from lightning from 1940-1949.

The precipitous decline in lightning fatalities is even more remarkable when you consider the U.S. population has markedly grown during the period of record.

Why so many tornado deaths?

Although tornado deaths spiked in 2011, they have trended downward over time, like lightning deaths, despite the growing population. And when you normalize tornado fatalities based on population, the downward trend is very sharp as shown in the graph below. No doubt better warnings and education have made us safer from tornadoes

Tornado deaths per million people in United States since late 1800s (NOAA)

2011 brought one of the two worst tornado outbreaks on record in the United States and three of the worst individual tornadoes. It also brought six tornadoes given the top rating of EF5.

Critically, several of year’s strongest tornadoes targeted major urban areas: Joplin, Birmingham, and Tuscaloosa.

Reducing future tornado risk

As I said in an earlier blog post, despite considerable progress in tornado forecasting and education, the tragic toll inflicted by 2011’s twisters demonstrates there is much work to be done to reduce the vulnerability of our population to tornadoes.

2011 is a clear indicator that tornadoes are a more difficult problem to confront than lightning.

Whereas it’s a relatively easy (and important) task to teach people “when thunder roars, go indoors” (NOAA’s lightning safety campaign slogan), we must face tackle many more complex challenges to protect a major city from a mile-wide EF-5 tornado with 225-250 mph winds.

NWS meteorologist NWS Gary Woodall, who led an assessment of the record-setting April 27 outbreak in the South, said many people died despite taking what they thought were proper measures:

...there is evidence that some people died while doing what they thought was the right thing. Many people who did not have a storm cellar or basement took shelter in an interior room of their home. However, a lot of the homes that were impacted were of wood-frame construction, and just could not withstand the battering of the strong and violent tornadoes.

He also mentioned some people who got the warning failed to act because they sought ground-truth of the tornado near their location:

Many members of the public our team interviewed said that, when a warning was issued, they looked for confirmation of the threat before taking action. This confirmation may have been from television or radio, from a trusted friend or family member, or in some cases, waiting to visually see the tornado. People also listened for the mention of specific towns or geographic features on which to base their action decision. With the fast speeds at which these storms were moving (typically 50+ mph), looking for secondary confirmation didn’t give folks a lot of time to get to safety.

Woodall said the assessment team put forward the following recommendations for improving tornado responses:

In the short-term, we should emphasize the low-probability but high-impact nature of tornadoes even more in our tornado safety campaigns. We should ensure that the databases that list locations in a warning area are as detailed as possible without becoming overwhelming. This should help the people who listen for specific geographic locations as a trigger to take action – the offices did this manually on April 27. We should aim for a more continuous flow of information between watches (which can be in effect for several hours before the storms arrive) and warnings (10-20 minutes before the storms arrive).


In the longer term, we really need to engage the social scientists in all aspects of our warning philosophy, content, and process. By embedding the social scientists into the process, our warning messages will have a better chance of bringing about the proper response from the public.

These are all sound recommendations. While there’s no telling if the 2012 tornado will match 2011 in its ferocity, we should take all steps to make the same progress in tornado safety that we have with lightning.