Hurricane Laura’s winds intensified from 75 mph to 140 mph in just 24 hours, taking the storm from a Category 1 to a Category 4.
Such rapid intensification, which drove Laura to make landfall in Louisiana with 150 mph winds last week, is difficult to forecast. But a team of researchers led by scientists at NASA’s Jet Propulsion Laboratory is hoping machine learning can help.
Future storms: Climate change might make Hurricane Laura’s 150 mph winds less rare
Rapid intensification occurs when a hurricane’s wind speeds increase by 35 mph (or more) within 24 hours. This is difficult to accurately predict because it’s dependant on both the environment outside and inside a hurricane. And on the inside, it’s difficult to measure a storm’s characteristics — such as how hard it’s raining or how quickly the air is moving vertically — and to determine which ones result in rapid intensification, according to a NASA news release.
To create a new forecast model, described in a paper published last week in the journal Geophysical Research Letters, the team of researchers sifted through years of satellite data and determined that rainfall rate in the dense wall of thunderstorms surrounding the storm’s eye is a good indicator of intensification. The harder it’s raining inside a hurricane, the more likely the storm is to intensify.
The researchers also noticed the ice water content of clouds within a hurricane and the temperature of air flowing away from the eye at the top of a hurricane factor into intensity changes.
Ultimately, the team added rainfall rate, ice water content and temperature of the air flowing away from the eye (called the outflow temperature) to predictors that National Hurricane Center uses in its operational model. Then the team came up with its own predictions using machine learning.
The researchers tested their model on storms from 2009 to 2014, and then they compared its performance with the National Hurricane Center’s operational forecast model for the same storms.
Expect more activity: An already busy hurricane season is about to kick into high gear
For hurricanes with winds that increased by at least 35 mph within 24 hours, the researchers’ model had a 60 percent higher probability of detecting the rapid intensification compared to the National Hurricane Center’s current operational forecast model. For hurricanes with winds that increased by at least 40 mph, the new model outperformed the operational one at detecting these events by 200 percent, according to the news release.
Hui Su, an atmospheric scientist at the Jet Propulsion Laboratory, was the lead author on the team’s paper. Su and her colleagues, including researchers from the National Hurricane Center, are testing their model this hurricane season and, in the future, plan to sift through satellite data for additional hurricane characteristics that could improve their machine learning model.
“It’s an important forecast to get right because of the potential for harm to people and property,” Su said in the release.