Back to Top

Researchers advance in tornado prediction

Researchers from New York put forth an improved method for tornado prediction in their paper, “Association of U.S. tornado occurrence with monthly environmental parameters.” The article is to be published in Geophysical Research Letters.

Certain weather phenomena like hurricanes may be predicted accurately and early enough to prepare or evacuate affected areas. However, there is often little to no notice about impending twisters. Warning sirens and alerts are not given until a tornado is directly observed or conditions are favorable. The researchers sought the ability to predict twisters days in advance rather than hours. Earlier warning would give affected areas time to prepare for severe weather conditions; ultimately, this would help to protect property and save lives.

The researchers studied the way that hurricane prediction techniques were developed, and applied a similar approach to tornadoes. Tornado records and advanced weather reports from 1979-2010 were obtained from several sources. The reports were adjusted for changes in reporting methods then the data were studied for environmental parameters that were related to tornadic activity. Weather patterns that were investigated included humidity, helical air currents, vertical winds, rain and precipitation related to severe weather.

By using a statistical test known as Poisson regression, the researchers determined which weather patterns would be most useful in predicting tornadoes. It was found that the key parameters were precipitation from atmospheric instability and the tendency of air to spin in updrafts.

A model was developed to make longer-term predictions of upcoming tornado frequency based on those weather patterns. The model was tested by using just the weather data from each month for a given area to predict the number of upcoming tornadoes likely in that region. The forecasted frequency of tornadoes was compared to the actual number to verify the accuracy.

With the exception of September and October, the model was found to be very effective for predicting tornado occurrence in comparison to current methods. The results were found to be most accurate in calculating likely tornado frequency in the late spring and early summer, the height of twister season.

Using this system, it should be possible to predict weeks in advance when and where a bad string of tornadoes is probable. While the model has been shown to be quite accurate against historical data, the next few months will bring the true test of effectiveness as current weather is used to predict upcoming tornadoes.