Application of Machine Learning Methods to Better Quantify Water-Level Anomalies in Annapolis, MD

Increased coastal development and rising sea levels can exacerbate coastal nuisance flooding, or high tide flooding, affecting infrastructure and local economies, especially in coastal communities currently at-or-around sea level. Coastal nuisance flooding is primarily a result of complex interactions between astronomical tides, local geomorphology, and meteorological forcing. Statistically robust methodologies to better understand and quantify the relationship between water level differences, local-scale geomorphology, and meteorological forcing across short-to-medium temporal scales are still lacking. Annapolis is home to the U.S. Naval Academy (USNA), and is located at the mouth of the Severn River, a tidal tributary of mesohaline Chesapeake Bay. Coastal nuisance flooding is already economically costly to the City of Annapolis and is predicted to become more frequent and severe as a result of sea level rise. This study utilized machine learning methods to develop a model to better predict water level anomalies (WLAs) in Annapolis, MD using locally-available water level and meteorological data. Results can be used to help local stakeholders develop better predictions for water-level anomalies in Annapolis, MD and mitigate against the economic and infrastructure losses resulting from coastal nuisance flooding. The ultimate goal is to develop a robust data science methodology to help better predict the frequency, timing, scope, and scale of nuisance flooding in coastal communities.