The Royal Institution for the Advancement of Learning/McGill University

Published purpose

Knowledge of ice conditions from freeze-up to break-up is critical for hydrological forecasting and monitoring, as well as for safe operation of ships in icy waters. However, forecasts of ice conditions are almost nonexistent for the St-Lawrence Seaway, a major shipping channel in Canada. This project will develop a forecast model for the (sub-) seasonal evolution of ice conditions in the St-Lawrence River based on Machine Learning. The purpose of these predictions is to aid stakeholders in decision-making and risk analysis for their activities. Ideally, the forecasts will have a lead-time of 30 to 60 days and will be updated daily. First, the model will focus on forecasting ice conditions in the St-Lawrence Seaway, since there is currently a lack of ice forecasts in this region. The model will then be extended to the Estuary/Gulf of the St-Lawrence River if the forecasts for the Seaway are shown to be skillful.