Conceptual and Modeling Approaches to Investigate Coastal Road Vulnerability to High Water Tables and Flooding

Abstract

Low-lying coastal areas face increasing risks from elevated groundwater levels and surface flooding due to sea-level rise (SLR), extreme precipitation, and storm surges. The objective of this research is to develop and apply a transferable framework to assess the vulnerability of coastal road infrastructure to shallow groundwater tables and flooding, using Alabama State Route 180 (AL-180) as a case study. The methodology integrates a data-driven empirical model and a physics-based distributed hydrologic model to evaluate sub-daily interactions between precipitation, tidal fluctuations, and groundwater dynamics. The empirical model estimates groundwater level (GWL) time series using simplified inputs such as rainfall, tidal elevation, surface elevation, and distance from tidal bodies. It was validated with field data from monitoring wells along AL-180. Performance metrics include Nash–Sutcliffe Efficiency (NSE) values of up to 0.73 and root mean squared errors (RMSE) as low as 0.06 m. This model was implemented along longer segments of the road using a Geographic Information System (GIS) platform at 100-meter intervals and used in conjunction with spatiotemporal analysis tools to detect long-term trends and identify vulnerable segments. Additionally, the Gridded Surface Subsurface Hydrologic Analysis (GSSHA) model was calibrated and validated in a coastal setting to simulate surface and groundwater processes during moderate and extreme events. Results show that combined surface-subsurface interactions have a significant impact on flood extent and pavement saturation, particularly during extreme rainfall events. Together, the modeling approaches identified road segments with high saturation persistence, informing locations for infrastructure maintenance and adaptation. This framework supports transportation agencies in developing resilient road systems and can be applied to other coastal corridors with minimal data requirements.