A6: Model Intercomparison and Analysis

A5 carefully benchmarks the GWF and collaborator land surface hydrologic models including MESH, VIC, WRF-Hydro (as applied by GLERL), HYPE, and GEM-Hydro amongst others to explore and improve modelling processes. Initial work with Lake Erie watershed modelling will be followed by the Great Lakes Basin and Saskatchewan-Nelson-Churchill River Basins (sub-basins selected in consultation with collaborators) in six years.

In Phase I of IMPC (Lake Erie and Great Lakes), models will be compared from both an ‘out-of-the-box’ perspective and driving a community effort to collectively identify common datasets, forcing datasets and model performance measures. The first phase developed frameworks to collaboratively make decisions and provide unified model setup information to ease the comparison of individual modeling decisions. Phase I also identified the large group of collaborators significantly exceeding the project team initially suggested. The project is based on three main objectives: comparing models for low-impact streamflow gauges, streamflow gauge stations most downstream to the lake(s), and comparing models based on their ability to simulate net basin supply.

In Phase II (Nelson-Churchill), we will apply the unified methodology for modelling intercomparison derived in Phase I, including time frames for model analysis, performance metrics, input data selection, and forcing data sets. The analysis will evaluate how and why models differ from a process-based perspective in unregulated basins, and evaluate methods of representing human impacts in the basin and their relative influence on model performance.

Area of focus:

• Lake Erie (GRIP-E)
• Great Lakes (GRIP-GL)
• Nelson-Chruchill (Nelson MiP)

Models in Use:

• MESH
• VIC
• WRF-Hydro
• HYPE
• GEM-Hydro
• Many more.

In the News