Simulation of snow water equivalent (SWE) using thermodynamic snow models in Quebec, Canada

International audienceSnow cover plays a key role in the climate system by influencing the transfer of energy and mass between the soil and the atmosphere. In particular, snow water equivalent (SWE) is of primary importance for climatological and hydrological processes and is a good indicator of climate variability and change. Efforts to quantify SWE over land from spaceborne passive microwave measurements have been conducted since the 1980s but a more suitable method has yet to be developped for hemispheric-scale studies, and tools such as snow thermodynamic models allow a better understanding of the snow cover and can potentially significantly improve existing snow products at the regional scale. In this study, the use of three snow models (SNOWPACK, CROCUS and SNTHERM) driven by local and reanalysis meteorological data for the simulation of SWE is investigated temporally through three winter seasons and spatially over intensively sampled sites across Northern Québec. Results show that the SWE simulations are in agreement with ground measurements through three complete winter seasons (2004–2005–2005–2006 and 2007–2008) in southern Québec, with higher error for 2007–2008. The correlation coefficients between measured and predicted SWE values ranged between 0.72 and 0.99 for the three models and three seasons evaluated in southern Québec. In subarctic regions, predicted SWE driven with the North American Regional Reanalysis (NARR) data fall within the range of measured regional variability. NARR data allow snow models to be used regionally, and this paper represents a first step for the regionalization of thermodynamic multi-layered snow models driven by reanalysis data for improved global SWE evolution retrievals