Approximately 1.4 million Californians live in areas that are at high risk of flooding each year, and climate change is only driving this number up. But flood risk can be a tricky concept for residents and policymakers alike to grasp. Just looking at a FEMA map or knowing that a home is built in a 100-year floodplain doesn't give people an intuitive understanding of how likely it is that a flood will occur - or what could happen when one does.
But what if it was possible for people to actually see what it would look like when their community floods, and how specific management decisions, like levee improvements, might change the picture?
That's the goal of an ongoing UC Berkeley-led project exploring how hydrodynamic digital twins can help California residents and policymakers better understand and manage flood risk.
"Digital twins are opening a whole new world on how we can think about risk," said Anna Serra-Llobet, a senior scientist at the Berkeley Center for Smart Infrastructure.
A digital twin is a dynamic, highly detailed virtual model of a real-world object, system or environment that can be updated in real time to mirror how its physical counterpart behaves. As a case study, the team is building a digital twin of the Central California farmworking communities of Watsonville and Pajaro. Located on either side of the Lower Pajaro River, the city of Watsonville and the unincorporated community of Pajaro have experienced five major floods since the Pajaro River levee system was completed in 1949. This includes a devastating flood in March 2023 that caused an estimated $300 million in damage.
After decades of delays, the levee system protecting Watsonville and Pajaro is finally getting a major upgrade. Berkeley researchers are working closely with the Pajaro Regional Flood Management Agency and other state, municipal and community organizations to construct a digital twin of the Lower Pajaro River Watershed, including the new levee system. If successful, this approach could be extended to other flood-prone regions of California. It will be the first digital twin for flood risk management created in the U.S. using special software called scenarify, developed by the Vienna Research Center for Visual Computing (VRVis GmbH).
VRVis GmbH
The project is led by Serra-Llobet and Civil Engineering Associate Professor Adda Athanasopoulos-Zekkos, who is also with the Center for Smart Infrastructure. Key collaborators include Berkeley Ph.D. candidate Gabriela Paredes, who is developing the digital twin model with VRVis GmbH researchers Jürgen Waser and Silvana Rauer-Zechmeister and Professor Kofi Boone of North Carolina State University, who brings expertise in community engagement and environmental justice.
A virtual version of reality
A digital twin is a virtual version of reality, where real-world objects, systems and processes are recreated as accurately as possible in a digital environment. Researchers at VRVis GmbH are already using digital twins in other countries such as Austria, Germany and Denmark to simulate potential future flooding scenarios. The Berkeley team would like to apply this digital twin technology in California and find ways to include climate risk, environmental considerations and social inequities into these simulations.
"When you look at a flood risk map, you only see a flat depiction of the hazard area, and this is what informs most of the decision-making around reducing risk," Serra-Llobet said. "With a digital twin, we can see potential flooding in three dimensions and see how processes change over time."
For example, a "setback levee" is a levee that is moved farther back from a river channel than a conventional levee, allowing pre-existing floodplains to be partially restored. Setback levees are growing in popularity because they reduce flood risk while, if well designed, also providing the possibility to integrate ecological benefits, including boosting groundwater recharge, generating new wetland habitats and opening space for new parks and recreation areas. However, they can also have drawbacks: Moving a levee back from a river might displace an established residential community that may already be socially or economically vulnerable.
A digital twin could potentially include all of these factors and more, making the pros and cons of any potential scenario more explicit.
"We have flood maps and we have social vulnerability maps, but it can be really difficult for local communities to understand what they mean and translate them into action," Serra-Llobet said. "We want to investigate how we can use digital twins now and then get input from all our team partners and community members to see how we can improve this software and better integrate all these topics."
When paired with extensive sensing systems, digital twins can also be used to monitor levees and other infrastructure as they age and undergo repeated stressors.
"We want to prolong the life of our infrastructure as long as possible, but at the same time it is experiencing destructive events that will only occur more often with climate change," said Kenichi Soga, a Distinguished Professor of Civil and Environmental Engineering at Berkeley and director of the Berkeley Center for Smart Infrastructure. "This is where digital twinning and monitoring can be very helpful, because you can assess what's going on and really understand the actual (infrastructure) behavior."
Kara Manke/UC Berkeley
Waking up to a flooded home
Shortly after midnight on March 11, 2023, the Pajaro River - engorged from weeks of heavy rainfall in Central California - breached its levees approximately three miles east of the town of Pajaro.
Water quickly rushed across the adjacent strawberry fields, and by 3 a.m. the flooding had started to enter people's homes.
"People got caught in the middle of the night waking up to water in their house," said Mark Strudley, executive director of the Pajaro Regional Flood Management Agency.
By sunrise, the entire community of 2,800 people was submerged in floodwaters, leaving hundreds in need of rescue and impacting more than 200 homes. The flooding also contaminated strawberry farms, delaying the growing season and dealing a massive blow to the local economy.
Though Congress authorized new levees for the Pajaro River in 1966, the project has been continually deprioritized because the property values in the area are relatively low in comparison to what it would cost to protect against flooding, giving the project a lower cost-benefit ratio under the system used to rate these projects. The Pajaro Regional Flood Management Agency was formed in 2021 to help overcome this and other administrative hurdles, and construction on a new levee system finally began in 2025. The new system will provide 100-year flood protection for the city of Watsonville and town of Pajaro as well as a mix of 100-year and 25-year flood protection for other unincorporated areas nearby. The project includes setback levees to increase flow capacity and restore floodplains and wetlands.
Serra-Llobet and the team chose the Pajaro River as a case study in part because of the Berkeley team's pre-existing expertise and connections to the community. The region also has a complex social and geographical history that makes it an interesting example of how equity and community engagement can be better integrated into flood risk management.
"We are still thinking about risk as a map, but risk is a process," Serra-Llobet said. "Climate change is changing precipitation and temperature patterns, bringing more intense atmospheric rivers in California. Changes to ecosystems are impacting the way the flood moves through the landscape. We have aging infrastructure and increasing urbanization that is changing hydrological patterns. We hope that digital twins can help us to put all these systems together."
