An Early Tornado Warning System

Voiceover

Presenter Information(s)

Anirudh Rao

Abstract or Description

Each year, on an average 1000 Tornadoes hit the United States resulting in 1500 injuries and 100 deaths. Almost 75% of tornadoes in the world happen in the central USA. On average, each tornado causes about $500,000 of losses resulting in a total loss of $500 million each year.

In addition, the tornado alley is moving east where it is densely populated, and this problem is worsening due to global warming. The 2023 early tornado season was a perfect example of this trend, where multiple tornadoes touched down east-wards as compared to the regular tornado alley. With the Gulf of Mexico warming, due to climate change, tornadoes are more and more expected in the Southeast and Midwest United States.

The current average lead-time for tornado warnings is 13 minutes. Today’s warnings are issued only when tornadoes are spotted on the doppler radar screen or when there are some eyewitnesses. Both offer very less time to save life and property. There is a big need to increase Tornado warning times.

Research in the last 20 years and recent test results have clearly established that tornadoes produce low frequency sounds or infrasound, of 0.5 to 20 Hz that can travel for miles. My project is about engineering a solution based on infrasound detection, combined with other environmental factors that have historically accompanied a tornado to increase warning times. The solution is based on a combination of sensors such as sensitive infrasound detectors, pressure, temperature and altitude, carried by autonomous drones. These drones are continuously monitoring in all directions, can be self-charging and launched from a base-station. 

Further testing with local weather stations and designated areas during the tornado season is needed to confirm the accuracy of the solution and calibrate with doppler radar results. In future, the solution can potentially also compare with existing data models using data science techniques, to increase accuracy of prediction.