When Jennifer Irish, a civil engineering professor at Virginia Tech, looks at a hurricane heading for the coast, she sees more than just how high the water might rise. She cares about how long that water sticks around.
"The timing and duration of storm surge are just as critical as its height when it comes to coastal destruction," Irish said. "A surge that lingers for days causes severe, prolonged erosion on barrier islands and keeps critical evacuation routes submerged long after a storm passes."
That insight guided Irish and three colleagues at Virginia Tech as they set out to understand storm surge behavior in new ways. Their research, published in the journal Coastal Engineering, analyzed more than 1,000 storm surge events from 62 named hurricanes that hit the United States between 2003 and 2022. Using a machine learning technique called k-means clustering, the team grouped the events based on how flooding evolved over time.
What they found surprised them: storm surge doesn't behave randomly. It follows eight predictable patterns. Some storms push water onshore quickly, then it retreats gradually. Others build slowly and stay elevated for days. The patterns repeat across different storms and coastlines.
"These recurring patterns provide a clearer picture of how storm surge evolves across different coastlines and storm conditions," said David Muñoz, an assistant professor who worked on the study.
The research also revealed that hurricane category does not tell the whole story. A Category 3 storm and a Category 3 storm can produce very different flooding timelines depending on their size, forward speed, direction, wind patterns, and the shape of the sea floor near the coast.
"Two storms with similar wind speeds can produce very different flooding timelines," said Atefeh Alipour, the doctoral student who led the research.
The Gulf Coast showed the most variety in how storm surge behaves, partly because its shallow ocean floor amplifies waves and because hurricanes hit there frequently. The Atlantic Coast had fewer overall patterns, but those patterns played out differently depending on which part of the coastline faced the storm.
For coastal communities, these findings could reshape how engineers design roads, bridges, and buildings. Instead of planning only for peak water height, planners could now design infrastructure that holds up against the specific timing and duration of flooding in their region. Emergency managers could also use the patterns to improve evacuation routes and shelter planning.
"As climate change contributes to rising sea levels and more intense tropical cyclones, understanding how storm surge evolves is more important than ever," said Robert Weiss, a geosciences professor on the team. "This framework could improve flood prediction and support better planning and decision-making to make coastal communities more resilient."