New technique can give vulnerable communities more time to prepare
The rapid intensification of a hurricane ranks among the most dangerous weather events on Earth. Storms such as Hurricane Michael, which crashed ashore the Florida Panhandle with winds of 160 mph, and Hurricane Otis, the costliest hurricane to strike Mexico's Pacific coast, caught communities unaware when they intensified at a rate that computer models failed to anticipate.
Now a research team led by scientists at the U.S. National Science Foundation National Center for Atmospheric Research (NSF NCAR) has developed a method to predict which storms are likely to intensify rapidly as far as five days in advance — potentially providing local officials with vital time to make lifesaving preparations.
The new method focuses on a statistical approach that analyzes how the rapid intensification of past storms in different locations was affected by a number of variables, ranging from sea surface temperatures to conditions high in the atmosphere. It then uses those variables to calculate the probability of a current storm rapidly intensifying.
The new technique builds on a similar method that has helped the National Hurricane Center better anticipate the rapid intensification of storms from one day to as much as three days in advance.
"The advantage of this new technique is to provide a tool to forecast rapid intensification even further out," said NSF NCAR scientist Christopher Rozoff, who is leading the research. "This can give needed time for those in harm's way to board up their windows, evacuate, or take other steps to protect lives and property."
The method is described in a new paper in Weather and Forecasting, a publication of the American Meteorological Society. The paper was co-authored by scientists at the University at Albany, State University of New York; and the Community Programs of the University Corporation for Atmospheric Research. Funding came from NSF and NOAA.
Prediction challenges
Rapid intensification is generally defined as occurring when hurricane winds increase by 30 knots (about 35 mph) in a 24-hour period. This can have lethal consequences, especially when a tropical storm or minimal hurricane suddenly gains intensity as it nears shore. Hurricane Helene, which needed just two days to grow from a relatively weak tropical storm into a major Category 4 hurricane at its Florida landfall in 2024, was blamed for about 250 deaths across several southern states. Hurricane Harvey, which rapidly intensified as it neared the coast of Texas in 2017, submerged parts of the Houston area with unprecedented rains and caused an astonishing $160 billion in damage.
Although hurricane forecasts have improved significantly over the last several decades, meteorologists still struggle to predict rapid intensification. When Hurricane Michael became a tropical storm in the Gulf in 2018, for example, forecasters expected its winds would top out at no more than 70mph. Instead, partly because upper-atmospheric conditions did not evolve as anticipated, Michael became a Category 5 storm that pulverized parts of Florida's Panhandle and caused an estimated $6 billion in damage to fighter jets at Tyndall Air Force Base.
The new method by Rozoff and his colleagues would not address all of the challenges in predicting rapid intensification. But it's intended as a tool that, in addition to computer models, can help meteorologists better anticipate the probability of a storm rapidly intensifying.
To create it, the research team focused on hurricane forecasts in the Atlantic and eastern Pacific that were generated from 2019 to 2024 by NOAA's Global Ensemble Forecast System, which is a leading weather model. The scientists applied a statistical technique known as a logistic regression model to the forecasts, systematically analyzing 55 variables in both the structure of the storm and the larger environment to determine which were most important as predictors of rapid intensification.
The resulting method performed well in evaluations with forecast models. For Atlantic hurricanes, it helped indicate the probability of rapid intensification as much as five days in advance, although that dropped to three days in advance for eastern Pacific hurricanes.
Furthermore, the method was useful in showing which of a hurricane's forecasted tracks was most likely to result in rapid intensification. This can enable forecasters to anticipate that, if a storm tracks in a certain direction, it has a far greater probability of rapid intensification than if it tracks in a different direction.
Future research can further improve the method, both by analyzing hurricane patterns over additional years and by testing it with additional forecast models, Rozoff and his co-authors wrote in the paper.
"There are a variety of ways to strengthen these predictions," Rozoff said. "The goal is to give forecasters enough information so they have confidence days in advance of when a hurricane is most likely to rapidly intensify."
About the paper
Title: An Ensemble-Based Tropical Cyclone Rapid Intensification Prediction Tool for
Extended Lead Times
Authors: Christopher M. Rozoff, Jonathan L. Vigh, Eric A. Hendricks, Brittany Freeman, and Paul A. Kucera
Journal: Weather and Forecasting
This material is based upon work supported by the NSF National Center for Atmospheric Research, a major facility sponsored by the U.S. National Science Foundation and managed by the University Corporation for Atmospheric Research. Any opinions, findings and conclusions or recommendations expressed in this material do not necessarily reflect the views of NSF.
