Engineering Resilience Before Disaster Strikes

As climate change intensifies storms, floods, droughts, wildfires, and coastal erosion, engineers are being asked to solve a difficult problem: How can communities prepare for disasters that are becoming more frequent, more severe, and harder to predict?

For Farshid Vahedifard, professor and Louis Berger Chair in Civil and Environmental Engineering at the School of Engineering, the answer begins with data collection about a wide range of issues, including levee performance and deterioration, wave and tidal action, wildfires, and soil conditions. He then uses that data to create computer models to predict the probability of extreme events, and how they may affect both infrastructure and the natural landscape.

“We are trying to predict and quantify the potential impact of extreme events,” Vahedifard said, “and then work with communities, end users, and decision makers to develop practical solutions that reduce risk and strengthen resilience before disasters occur.”

Protecting Coastal Cliffs

On Martha’s Vineyard, just south of Cape Cod, Vahedifard is leading a multidisciplinary team of engineers, environmental scientists, and community-engaged researchers from Tufts working with the Wampanoag Tribe to study erosion at the Aquinnah Cliffs. 

The tribe recently received a grant from the Massachusetts Office of Coastal Zone Management to evaluate the effects of storms and sea level rise, and identify ways to stabilize the cliffs against erosion.

“Those cliffs are not just a geographic feature,” said Vahedifard. “They have a very unique cultural importance.” 

Tribal members have used clay from the site for pottery over generations, and the cliffs are tied to tribal lore and identity. That means any proposed engineering solution must respect the tribe’s priorities to preserve the cliffs’ spiritual and cultural significance. 

Vahedifard is working with Elaine Donnelly, director of the Tisch College Community Research Center at the Jonathan M. Tisch College of Civic Life, to engage the Wampanoag community in the process.

On the technical side, Vahedifard and his research team are developing a physics-based model together with remote sensing to understand how the cliffs have changed over time and how they may respond under future scenarios of sea level rise, wave action, rainfall, and vegetation change. 

The team is exploring nature-based solutions such as vegetation that could dissipate wave energy and reduce erosion without adding hard infrastructure like concrete barriers that would alter the natural character of the landscape. 

Predicting Extreme Wildfires

Vahedifard has also led a multidisciplinary team of researchers from several universities investigating the impacts of wildfires on both natural and engineered systems. They examined post-wildfire landslides, debris flows, and other cascading hazards that can threaten communities long after a fire has been extinguished. 

The research aims to improve understanding of both wildfire behavior and its downstream consequences, ultimately supporting more effective preparedness and resilience planning.

As part of this effort, Vahedifard and his team recently investigated one of the most vexing problems in wildfire preparation—the difficulty in predicting unusually large and devastating fires, such as those that struck Hawaii in 2023 and California in 2025. Although these extreme events represent only a small fraction of all wildfires, they account for a disproportionate share of the damage and losses. 

“The goal of the project is to better understand risk factors leading to, and cascading events following, these wildfires and how we can mitigate their impact on communities, particularly vulnerable communities,” he said.

Vahedifard and colleagues analyzed more than 30,000 fires in the United States over roughly four decades. On a graph, most fires would cluster around the middle of a bell-shaped distribution, with smaller and larger events becoming progressively less common. However, the researchers found that a small number of exceptionally large fires form an extended “tail” on the distribution curve that represents the most catastrophic events. 

Using advanced statistical models, the researchers developed a nationwide assessment covering 125 ecologically distinct regions and identified areas that are particularly susceptible to extreme wildfire events. 

Emergency managers, forestry agencies, land-use planners, and other public officials can use this information to prepare for and reduce risk where extreme fires are most likely to occur. Such measures may include vegetation management, prescribed burning, clearing strips of land as firebreaks, enhanced public education and evacuation planning, and the adoption of more fire-resistant building materials and landscaping practices.

Digital Twins for Levees

Vahedifard has spent much of his career studying levees—the flood-protection systems that safeguard more than 23 million people in the United States along coastlines and major rivers. Many of these levees are aging, while at the same time facing increasing pressures from climate change, including more frequent flooding, sea-level rise, and extreme weather events.

The traditional approach to levee monitoring relies heavily on periodic visual inspections, which can be difficult to conduct across thousands of miles of infrastructure. “You cannot just put instruments along the entire levee system,” Vahedifard said. “It’s not like a bridge that is much easier to monitor.”

To address this challenge, his research team is working to create a “digital twin” for levee systems—a virtual representation that integrates field observations, remote sensing data, physics-based computer models, and artificial intelligence to continuously assess levee conditions and evolving risks. 

The digital twin can use information from well-monitored levee sections to improve predictions in areas where sensors and data are limited, helping engineers and agencies to better target maintenance and rehabilitation efforts, and help communities better prepare for flood emergencies.

For Vahedifard, engineering resilience means bringing together field data, remote sensing, and computer modeling to understand how the natural landscape and structures built by people respond to extreme events in a changing climate. This approach makes it possible to predict where risks are rising, identify practical ways to reduce them, and ensure that communities are part of deciding what protection should look like.