A lifeguard flag indicates dangerous swimming conditions.
Credit: dr_tr / flickr.

By Laura G. Shields

Being pummeled by a large wave can quickly ruin a perfect beach day, as it may require a trip to the hospital or even worse. But researchers in Delaware have developed a new method that could predict when wave-induced injuries are most likely to occur, potentially reducing incidences of these injuries to unsuspecting beachgoers along the state’s coast, according to scientists who presented the research last month at the 2017 American Geophysical Union Fall Meeting in New Orleans.

The researchers hope to provide Delaware beach patrols with a tool to better understand what environmental conditions and human behaviors increase the probability of injuries. If lifeguards can forecast when injuries are likely to occur, they can perhaps increase staffing at those times and spread awareness among beachgoers, said Matthew Doelp, a coastal engineer at the University of Delaware who presented the new research.

Since 2010, Beebe Healthcare in Lewes, Delaware – the only trauma center in the area – has collected summertime data on surf zone injuries occurring at five public beaches along 24 miles of the state’s Atlantic coastline. This local hospital sees 130 – 450 injuries each summer, which range in severity from minor fractures and dislocations to more severe spinal injuries, according to the researchers. These wave-driven spinal injuries, as well as the six fatalities over the past eight years, are their primary concern, Doelp said.

Researchers used RTK GPS equipment to profile the beaches daily.
Credit: Matthew Doelp.

“We’re one of the first studies that are doing an in-depth recording of all these types of injuries,” he said. Past studies have looked at injuries from surfing or rip currents, but wave-driven impacts are more frequent than rip current rescues along the Delaware coast, Doelp said. The majority of incidents occur to waders, who are knocked over by waves when their backs are to the ocean. Tourists, who are often unfamiliar with the dangerous shore break, are six to seven times more likely to be injured during low-risk wading, whereas locals are injured by higher risk activities, such as surfing.

In addition, males are twice as likely to be injured as females, according to the researchers. Doelp was surprised by how consistent the demographics of the injuries are from year to year, and this pattern of injury characteristics inspired the researchers to see if they could predict when they would occur.

In addition to injury data provided by the hospital, the researchers analyzed how various environmental conditions at Bethany Beach, Cape Henlopen State Park, Delaware Seashore State Park, Dewey Beach and Rehoboth Beach change hourly. They collected data on nearshore wave height, peak wave period and direction, air temperature, wind speed, precipitation, solar radiation and wind direction. They also counted the bathers once daily and extrapolated that information to estimate the number of people in the water for each hour observed by the lifeguards.

The researchers inputted the data from these parameters into a computer model to see if knowing these environmental conditions and the number of bathers could help them predict the probability of an injury occurring. They fed the data collected from previous years into the model and then tested its forecasting performance on data collected in 2017.

They found their injury likelihood model predicted occurrences well compared to prior predictions. The model explained 79 percent of the observed data’s variation.

This Bayesian network absorbs most of the uncertainty of the human decision-making risk into the other parameters, Doelp said. The human behavior component will likely have more error associated with it than the environmental parameters, but the trained model relies on probabilities and can make predictions without all relevant parameters, he explained. The more data he inputs into the model, the better it should do.

Doelp is considering developing an app to provide lifeguards with the data in real-time.

— Laura G. Shields is a science communication graduate student at UC Santa Cruz. Follow her on twitter at @LauraGShields and read more of her work at lauragshields.com.