Field Testing and Comparison of Four Snowmelt Model Approaches in the Quesnel Highlands, Interior British Columbia

In an attempt to improve the performance of the basic temperature-index (TI) snowmelt model, modifications, including the addition of energy balance terms, have been introduced. One of the more popular modifications has been the inclusion of a shortwave radiation term or proxy term. We compared the performance of a selection of temperature-index snowmelt models at a partly forested site in the interior of British Columbia, Canada. A basic TI model, a TI model with slope, aspect, and forest cover correction factors (HBV-EC, CHC 2006), a TI model with a modeled potential incoming shortwave radiation term (Hock, 1999), and a TI model with a measured shortwave radiation term (Pellicciotti et al., 2005) were compared against snow course measurements on north and south aspect open and forested sites, total season melt on east and west aspect open and forested sites, and against daily automatic snow pillow melt data. Including the effects of slope and aspect and forest cover in some form in TI models does appear to improve melt modeling. The HBV-EC model performed better than the basic TI model and about the same or better than the two models that included solar radiation explicitly. There was little difference observed between the slightly varying radiation index models from Hock (1999) and Pellicciotti et al. (2005).