Conceptual data-driven approach for analyzing the vulnerability of coastal roadways to groundwater level changes

Coastal infrastructure, such as roadways, is particularly vulnerable to water stressors. An excellent example of such a vulnerable system is the State Route AL-180, located in Fort Morgan Peninsula, Alabama. While much attention has been devoted to studying the effects of extreme hydrological stressors on roadways, the effects of moderate rain events, tides, and elevated groundwater levels (GWLs) on the saturation of pavements in low-lying areas have been mostly ignored. GWL predictions typically apply analytical, numerical, or data-intensive empirical formulations. Typically, data-intensive formulations can be computationally intensive and involve numerous assumptions. In this effort, we propose a relatively simpler new data-driven methodology for groundwater level estimation in coastal roadways. The method is used to evaluate the impacts of pavement exposure to elevated shallow GWL for the AL-180 highway. The modeling framework is based on GWL response, which is influenced by rainfall, distance to tidal water bodies, tidal level fluctuations, and groundwater storage characteristics. Water level data collected from two shallow wells along AL-180 were used to calibrate the proposed model, and a third well dataset was used for validation. Results show that the Nash-Sutcliffe efficiency (NSE) varied from 0.65 to 0.73 when the first two locations were considered, and for the validation location, the NSE was 0.56. The projected GWL data also indicated sections of AL-180 may be fully saturated 100 % of the time because of sea level rise. The proposed modeling framework is a good tool to guide decision-making processes for coastal roadway management. However, the results are constrained by simplified assumptions, such as uniform soil characteristics, no lateral flows, and limited spatial validation.