Towards an operational hydrological and hydrodynamic forecasting system for the Great Lakes and St. Lawrence River watershed at Environment Canada

Environment Canada's Canadian Meteorological Centre (CMC) is preparing to launch a new operational environmental forecasting system for the Great Lakes and St. Lawrence. These forecasts will be produced by a cascade of coupled numerical models. In the first version of the system, the High-Resolution Deterministic Prediction System (HRDPS), which is based on the GEM atmospheric model, will provide atmospheric and surface hourly forecasts for 48-h at a 2.5-km resolution, including surface runoff and groundwater recharge to a hydrological routing model (WATROUTE). The resulting estimates of streamflows will then be passed from the hydrological model to the ocean/lake model (NEMO), along with atmospheric input fields. In turn, NEMO will provide initial conditions of water surface temperature and marine ice cover to the weather forecasting model. Flows in the St. Lawrence River are obtained using the SHOP hydrodynamic prediction system, which relies on the H2D2 hydrodynamic model. We describe the modelling system and its products, present an evaluation of the water level forecasts in the Great Lakes with an experimental configuration of this system during the fall and winter of 2014-2015, discuss how it was used during the Toronto 2015 Pan-Am games, and discuss improvements expected to the system in the near future, including two-way coupling of the atmosphere and lake models, data assimilation of streamflow and water level observations, ensemble forecasting and applications in other watersheds.