Data assimilation of hyper-local water level sensors for real-time monitoring of coastal inundation: Fid-Bau Portal

Data assimilation of hyper-local water level sensors for real-time monitoring of coastal inundation

Son, Youngjun / Di Lorenzo, Emanuele / Park, Kyungmin / Wipperfurth, Spenser / Luo, Jian

Abstract As flood events become increasingly prevalent in coastal regions with sea level rise, multiple communities have deployed water level monitoring networks across estuaries in addition to existing tide gauges located primarily at immediate coasts. Due to the spatially-distributed nature of sensor deployments, however, water level data are only available at specific sensor locations during the time of monitoring. As a result, an information gap on water levels exists along estuarine channels outside of active monitoring locations. To fill such a gap, this study presents a physics-based empirical modeling approach to assimilate coastal water levels using observations from hyper-local water level sensors. We implement an Objective Analysis (OA) procedure for sensor observation datasets obtained from the Smart Sea Level Sensors project along the U.S. Georgia coasts, based on spatial covariance structures of water levels that are extracted from high-resolution coastal-ocean hydrodynamic simulations. The approach is validated using simulated water levels, which provide basis functions for spatial covariance information. Additionally, the implementation is validated using actual water levels from active monitoring stations. The assimilated results of water levels are compared to those obtained by the same OA procedure but with a commonly-used Gaussian covariance function, which lacks prior knowledge of spatial covariance structures. To demonstrate the capability of the assimilation approach, we extend its application to a hurricane event, during which other dynamic processes may be relevant to variability in coastal water levels. Overall, the presented approach provides an accurate and efficient estimation of estuarine water levels along channels, which can support community officials to promptly identify localized flood threats to critical infrastructure systems in coastal regions.

Highlights • Objective Analysis is implemented to reconstruct water levels from sensor observations. • Spatial covariance statistics are derived from coastal-ocean hydrodynamic model simulations. • Variability in water levels is better captured by combining spatial covariance information. • Real-time assimilation can provide situational awareness of floods in coastal communities.