A curb just six inches higher than the road can change where a Baltimore rainstorm sends its water. That small rise is one reason researchers spent three years working with residents to understand where the city floods and what could help.
Flooding in Baltimore has become more frequent and intense since records began in 1748, and researchers expect the problem to grow as the climate changes. Many older flood-control systems rely on concrete and other surfaces that keep rain from soaking into the ground. In crowded city neighborhoods, that can send water rushing along streets instead of helping it drain naturally. The Penn State-led team focused on flooding caused by rainfall, known as pluvial flooding, rather than river overflow or storm-driven high tides. Residents told researchers during city workshops that this kind of flooding was a major concern.
Over three years, the team asked Baltimore residents about their concerns, priorities and what kinds of fixes might work in their neighborhoods. Those conversations helped shape the flood model and the ways researchers tested possible responses. The ideas included planting rain gardens, which can help absorb rainwater, and cleaning storm inlets so water can enter drains. The researchers found that these community-designed steps could lower flood depths by as much as 55% in some locations.
Building a useful model was difficult. The ground in cities has tiny features that can redirect water, such as the edge between a curb and a road. Elevation measurements from a distance can capture large hills and valleys, but may miss those small changes. Flooding from rain can also begin and disappear within a few hours, leaving little time for measurement. The team had no street sensors to record water depths where flooding occurred.
To check whether the model matched real events, researchers instead used location-tagged social media posts and news images. They compared visible water levels with familiar heights, such as curbs. That approach offered a way to assess the model despite the lack of direct sensor records, though the researchers noted that sparse observations make validation challenging.
The findings, published in Sustainable Cities and Society, are part of the Urban Integrated Field Program, which brings researchers, city planners and residents together to study climate risks in American cities. Lead author Ava Spangler, a Penn State graduate student in geosciences, described the work as a framework in which community and technical experts shape both model development and adaptation plans through ongoing engagement. Co-author Antonia Hadjimichael, an assistant professor of geosciences, also pointed to the difficulty of measuring floods that can cause damage in the short time before they recede.
The Baltimore project links technical modeling with residents’ knowledge of streets and neighborhoods. Its results suggest that flood planning can test practical, locally chosen steps—not only large engineered barriers—and identify places where they may make the biggest difference. The researchers’ next challenge is to help cities use approaches that residents support and officials can carry out, while improving the data available to understand sudden street flooding.