Sensitivity of the Snowmelt Runoff Model to snow covered area and temperature inputs

Abstract

Snowpack is an important source of freshwater. Snowmelt runoff models provide a means to predict the timing and magnitude of spring snowmelt. This study assessed the sensitivity of the conceptual, degree-day Snowmelt Runoff Model (SRM) to snow covered area and temperature inputs in a small mountainous catchment in Utah, USA. It was found that snow cover products from the Moderate Resolution Imaging Spectroradiometer (MODIS) underestimate snow covered area during the second half of the melt season, leading to inadequate modeling of spring snowmelt with SRM (Nash-Sutcliffe Efficiency = 0.59; bias of −26.9%). Incorporation of ancillary snow covered area information provided by Landsat ETM+ imagery greatly improved streamflow simulations (Nash-Sutcliffe Efficiency = 0.92; bias of −0.01%). For the temperature input, SRM appeared more sensitive to the elevation and location of the temperature reference station than to the lapse rate scenario used to extrapolate temperatures throughout the watershed. These findings will be useful for snowmelt runoff research and water resource management.