The devastating tornadoes in Oklahoma and Arkansas this weekend was a sad way to start tornado season. Sensing equipment and forecasting models have been used in improved advanced warning systems to help people take shelter, however, major tornadoes are still pretty deadly. The recent 2011 tornado in Joplin, MO was one of the deadliest ever. The “Weather Forecasting Improvement Act of 2014,″ which passed the House in April 2014, seeks to address this need.
While there is room for improvement when it comes to advanced tornado warnings, the cost associated with the warning systems is somewhat controversial. Federal funding for disasters is essentially a zero-sum game, so high-profile disasters like tornadoes can use up a disproportionately high amount of the budget allocated to disaster research, leaving us vulnerable to other less news-worthy disasters. Since there are so few tornado-related fatalities every year compared to other weather disaster (e.g., heat waves), there isn’t much of a potential to save lives regardless of how good the advanced warning system becomes.
Meteorologist Eric Holthaus wrote a nice article on Slate about this issue [Link]
While it’s necessary to continue making progress on hurricane and tornado forecasts, it should definitely not be at the expense of funding to improve forecasts of lower-profile weather and climate disasters that, in aggregate, kill dozens of times more people per year, and are increasing. Essentially, the bill invests scarce funds in high-profile weather events at the expense of those that cause many more deaths. Boosted by human-caused climate change, heat waves now kill more people in the United States each year than hurricanes, tornadoes, lightning and floods combined, according to the CDC. And weather-related traffic accidents kill 10 times more than heat waves—more than 6,000 people per year. [see the first image below to visualize these magnitudes]
Track the bill here. The bill focuses on forecasting, but I am going to take this a step further and examine the entire warning system to see if there are other ways we could save lives. The ultimate goal is to keep people safe, and there are many ways to do so. I grew up with tornado sirens, which were great so long as you were awake. I was surprised to learn that this is generally the (see the bottom figure below). TV warnings are another old-fashioned way to warn people, which seemed to work back in the era when people watched network TV. Virginia does not have too many tornado sirens (although it had tornadoes!), and since I do not watch much TV, my family missed a couple of warnings. Later, Virginia Commonwealth University used the campus siren installed to warn us of active shooters on campus as a makeshift tornado alarm. In my opinion, that was terrific. They have had more tornadoes than shootings.
I later learned that I could get the best warnings from following meteorologists and weather nerds on twitter. I highly recommend following nearby National Weather Service offices on twitter to get high quality information in real-time (get started here). Weather apps can deliver warnings even if you aren’t actively using the app. I think these are all great options, but I am still sometimes blissfully unaware of serious weather despite being “prepared.” Being able to reach everyone including those who cannot be reached by the most far-reaching alerts in our shifting technology landscape is not a forecasting problem (e.g., the deaf cannot hear sirens). The point is that there are some serious challenges in ensuring that the warnings get to the people quickly. Doing so isn’t so much of a forecasting problem, unless the forecasts give much, much earlier warnings (think: hours instead of minutes).
Another problem is that I can always choose to ignore the warnings even if I get them. This will likely happen if there are too many weather warnings/false alarms (The Weather App That Cried Wolf). I blogged about this issue here.
The OR tie in for this post may be somewhat dubious, but the connection here is that these issues cut to systems thinking and tradeoffs. Plus, I’m a daughter of the Midwest and will always be somewhat obsessed with weather and tornadoes. Your thoughts on optimal ways to keep people safe (broadly speaking) from tornadoes are welcome.
From “Investing More Money in Tornado Research Would Be a Disaster” by meteorologist Eric Holthaus. Click through for more information.
Related factoid: most of the world is not at an increased risk of tornadoes. So tornado prediction and preparedness is mostly a United States problem
Parts of the world at an increased risk of tornadoes are in red. Courtesy of NOAA. Click through for more info.
Another factoid: tornadoes are far more likely to occur in the late afternoon and early evening than overnight. I was shocked to learn this.
Tornado likelihood across the United States. Click through for more information.
Related posts:
Punk Rock Operations Research 
October 16th, 2016 at 8:58 am
Great post.
I wonder if you had the same comparison but instead of fatality count the metric is total cost of the damage.
I don’t have enough numbers, but I guess that even though fatalities are low, these extreme weather events have a much higher material cost for damage than others. Which can be a reason for allocating more funds to predict/mitigate their damage.