LAUSR

Abstract. From June 19 to June 22, 2013, intense rainfall and concurrent snowmelt led to devastating floods in the Canadian Rockies, foothills and downstream areas of southern Alberta and southeastern British Columbia. The complexity of the topography in the mountain headwaters, presence of snow at high elevations and other factors challenged hydrological forecasting of this extreme event. In this study, the ability of the Global Environmental Multi-scale hydrological modelling platform (GEM-Hydro), running at a 1-km grid spacing, to simulate hydrometeorological conditions in several Alberta rivers during this event is assessed. Four quantitative precipitation estimation (QPE) products were generated using the Canadian Precipitation Analysis (CaPA) system by varying (i) station density and (ii) horizontal resolutions (10, 2.5 and 1 km) of the GEM precipitation background. CaPA at 2.5 and 1 km including all available stations in the headwaters provided the most accurate estimation of intensity and total amount of precipitation during the flooding event. Using these products to drive GEM-Hydro, it is shown that QPE accuracy dominates the ability to predict flood volumes. Initial snow conditions also represent a large additional source of uncertainty. Default GEM-Hydro simulations starting with almost no snowpack at high-elevations led to a systematic underestimation of flood volume and peak flow. Gridded estimates of snow water equivalent from the Snow Data Assimilation System (SNODAS) were also considered. They led to contrasting abilities to simulate flood discharge volumes and a consistent overestimation in the headwater catchments, illustrating the strong need for a reference snow product in the mountains of Western Canada. Finally, GEM-Hydro did not predict peak flow timing and hydrograph shape well. Model sensitivity tests show that it could be improved by adjusting the Manning coefficients, suggesting the need to revisit the routing parameters. There may be a need to include water management effects on flood hydrographs as well. These results will guide the development of GEM-Hydro as a hydrological forecasting system in Western Canada.