Impact of Ship Characteristics on Wave Dynamics and Sediment Resuspension in the Upper St. Lawrence River: A Field Data Analysis

This study examines ship-wave–driven nearshore turbidity responses in the Upper St. Lawrence River by combining empirical analysis with sequence-based modeling designed to test short-lived temporal dependence. Building on previous field observations, it quantifies associations between ship-generated wave characteristics, near-bed orbital velocity, wave-induced shear stress, and turbidity adjacent to a deep-draft navigation corridor. A time-delayed multivariate regression model and a Bidirectional Long Short-Term Memory (Bi-LSTM) neural network were applied to 2 Hz wave, turbidity, and vessel traffic data sets to examine passage-scale relationships and response timing. Higher vessel speeds and larger primary wave heights were associated with increased near-bed shear stress and turbidity, with the strongest responses observed at shallow, fine-grained nearshore sites. Feature importance rankings from the Bi-LSTM indicate that instantaneous near-bed hydrodynamic forcing dominates turbidity prediction, while lagged variables contribute only limited additional skill, suggesting short-lived carryover between closely spaced passages. These results compare the relative influence of wave-component structure, near-bed forcing proxies, and vessel descriptors on ship-wave impacts in a confined inland navigation corridor. Limitations include the seasonal sampling window, the focus on two nearshore stations, and the omission of explicit wind and discharge forcing, which define priorities for future refinement.