Predicting Interannual Sea Level Variations Along the U.S. East Coast Using Machine Learning and Climate Indicators

Coastal sea level changes have profound impacts on coastal ecosystems, infrastructure, and communities. Interannual sea level variations along the U.S. East Coast are influenced by a combination of dynamic and thermodynamic processes, including local wind forcing, Gulf Stream variability, regional ocean circulation changes, and thermosteric contributions. These processes are interconnected and strongly modulated by large-scale climate modes such as the North Atlantic Oscillation (NAO), El Niño-Southern Oscillation (ENSO), and Atlantic Multi-decadal Oscillation (AMO). This study leverages machine-learning-based predictive models to quantify and forecast interannual sea level variability by integrating diverse climate indicators. By incorporating indices of large-scale climate modes alongside local and regional oceanographic parameters, the model quantifies the relative contributions of each factor and identifies the dominant processes driving observed variability. The results demonstrate the potential of machine-learning approaches to capture complex nonlinear relationships between climate modes and regional sea level changes. NAO-driven atmospheric forcing and ENSO-related ocean-atmosphere interactions emerge as key predictors, with the models successfully replicating observed variability along different sections of the U.S. East Coast. The findings highlight the importance of integrating large-scale climate dynamics into regional sea level prediction frameworks and suggest new opportunities for improving forecast accuracy at interannual timescales.