News•October 17, 2012

Pacific Ocean Patterns Tied to Violent Tornado Outbreaks

The key to forecasting the severity of a tornado season may lie thousands of miles away from America’s “Tornado Alley,” in the wide expanse of the tropical Pacific Ocean, according to a new study released on Monday. The research, published in the Journal of Climate, demonstrates a tie between a particular pattern of sea surface temperatures in the tropical Pacific and large outbreaks of violent tornadoes — twisters ranked EF-3 and above on the Enhanced Fujita Scale — in the lower 48 states during the months of April and May.

The study could help weather forecasters develop methods to make seasonal projections of the likelihood of major tornado outbreaks.

Tornado damage from the deadly May 22, 2011 tornado in Joplin, Mo.
Credit: Flickr/FEMA.

Researchers at the University of Miami and the National Oceanic and Atmospheric Administration (NOAA) were motivated to conduct the study after the deadly 2011 tornado season, in which 538 people died, most of them during the months of April and May. Violent tornadoes tore through communities like Tuscaloosa and Birmingham, Ala., as well as Joplin, Mo., where 161 people perished in an EF-5 tornado on May 22.

The researchers used both observations and computer simulations to identify long-term climate signals that might help provide predictability for extreme tornado outbreaks. The ocean configuration they zeroed in on is known as “Trans-Niño,” since it occurs during the transition phase as an El Niño or a La Niña event begins or comes to an end.

In its “positive phase,” Trans-Niño features cooler-than-average sea surface temperatures in the central tropical Pacific, and warmer-than-average sea surface temperatures in the eastern tropical Pacific.

This configuration was present in seven of the top 10 extreme tornado outbreak years during the 1950-to-2010 period, the study found, including 1974, which featured the so-called “Super Outbreak” of tornadoes. Importantly, this ocean pattern was also present in 2011.

“The number of intense tornadoes during these worst outbreak years, those rated as F-3 to F-5 on the Fujita scale, was nearly double the number in other years,” said lead author Sang-Ki Lee or the University of Miami and NOAA.

U.S. tornado tracks during 1950-2011.

Credit: NOAA/Storm Prediction Center

The reason the sea surface temperature pattern is important is because it influences weather patterns downstream, particularly across the contiguous U.S. The positive Trans-Niño pattern, the study found, favors a jet stream that enhances clashes between cold, dry air from Canada and warm, humid air from the Gulf of Mexico. It also favors a more abundant supply of atmospheric wind shear, thanks to a strengthened southeasterly airflow over the Southeast and Central states. Wind shear, which occurs when winds vary in speed or direction with height, is a critical ingredient in tornado formation.

“This spatial pattern of sea surface temperature in the tropical Pacific produces the large-scale atmospheric patterns conducive to tornado outbreaks over the Central U.S.,” Lee said via email.

The ferocity of the 2011 tornado season was not anticipated far in advance, largely because weather forecasters lack reliable methods of making predictions about the severity of tornado seasons as well as which parts of so-called “Tornado Alley” might be most at risk (meteorologists have improved their ability to detect individual tornadoes and warn people in their path, however).

This study could help change that, said Harold Brooks, a tornado researcher at NOAA’s National Severe Storms Laboratory in Norman, Okla. But scientists are still years away from being able to provide skillful seasonal tornado forecasts.

“I'm not sure how far away we are — maybe 5 years, maybe 25,” Brooks said in an email conversation. “There are several groups doing research right now that's relevant. One of the big problems is that most of the things that have been found still only explain a small fraction of the variance [in the number of very intense tornadoes from one year to the next]. Fundamentally, the question is what should we be forecasting? Should we forecast total number, path length, number and severity of large outbreaks, or something else? This study suggests that we might have hope of doing something with the outbreak question.”

Lee also said that predicting the likelihood of large tornado outbreaks well in advance may rest, in part, on scientists’ ability to accurately predict the evolution of sea surface temperatures in the Pacific, an area where gains have been made, but where forecast skill remains poor, particularly during the spring months.