Impact of assimilating snow cover observations for streamflow simulation and forecasting in Québec

Cantet et al. (2019) recently developed an observation product for snow distribution over southern Québec. It consists of 10 km x 10 km snow water equivalent grids produced by the HYDROTEL snow accumulation and ablation module, in which manual snow water equivalent measurements were assimilated using a spatially correlated particle filter.

Snow is a crucial hydrological variable for hydrological simulation and forecasting, in particular for anticipating the magnitude and timing of spring freshet. It is possible that assimilation by direct insertion of the new snow water equivalent grids in the HYDROTEL model will improve streamflow simulation and forecasting in Québec.

Source : Cantet et al (2019). Journal of hydrometeorology

• Verify whether the assimilation of snow water equivalent grids produced by Cantet et al. (2019) in the HYDROTEL model could improve hydrological simulations at gauged and ungauged sites in Québec as well as hydrological forecasting;

• Evaluate the added value of snow data assimilation in historical streamflow simulation (1970-2018) with HYDROTEL.

• Assimilate the Cantet et al. (2019) snow water equivalent maps in HYDROTEL by direct insertion:

  • Comparison with assimilation of SNODAS maps and with “open loop” simulation (i.e., without assimilation);

  • Verification of performance at gauged and ungauged sites (leave-one-out approach);

• Compare different temporal frequencies for assimilation;

• Conduct ESP/Reverse-ESP runs to compare the influence of snow data assimilation with the influence of dynamic weather forecasting on hydrological forecasts.

The project will lead to the production of new historical streamflow simulations (1970-2018) with HYDROTEL by integrating snow data assimilation. It will also contribute to a better understanding of the relative contributions of snow data assimilation and dynamic weather forecasting on hydrological forecast performance