PNNL

Abstract

Low-lying coastal urban cities are at significant risk of flooding, which can arise from a combination of factors, including river flooding (fluvial), storm surges and high tides (coastal), and intense local rainfall (pluvial). Typically, urban flood risks are assessed using design flood scenarios for specified return periods (e.g., a 100-year flood event). This approach, while computationally efficient, often oversimplifies the complex nature of multi-driver flood processes. Additionally, it largely overlooks the frequent but less intense floods, which are essential for flood risk management. To address these limitations, this study simulates over 200 flood events spanning 35 years (1985-2019) in Philadelphia, PA, a coastal city exemplifying multi-driver flood risks. The DHSVM-FVCOM-RIFT modeling framework was used to simulate urban flooding at a 10-m resolution, explicitly representing the interaction between fluvial, pluvial, and coastal processes. From these simulations, we identified flood risk hotspots (e.g., South and West Philadelphia) and the drivers of flooding in these areas. Analysis of 70 major flood events suggested that 44.3\% were compound events, primarily driven by fluvial-surge and fluvial-pluvial factors. Notably, 77.1\% of these events involved fluvial flooding, either as a single flood driver or in combination with other drivers, underscoring the dominant role of riverine processes in the city's flood risk. Additionally, we found that flood hazard analysis based on extreme design flood events could significantly underestimate the areas at risk of flooding. For instance, using Hurricane Irene (2011) as a proxy, the analysis captured only 44.2\% of at-risk areas. This highlights the importance of integrating a full spectrum of flood scenarios into coastal urban flood risk assessments to better inform the design of a flood-resilient landscape.