Snow modeling is a crucial part of spring streamflow forecasting throughout the western United States and the National Weather Service (NWS) SNOW17 model plays a primary role in determining the quality of the spring snowmelt predictions. Improvements made to the SNOW17 and/or NWS snow simulation methods could positively impact the NWS forecasts in this region. This study compares the SNOW17 to a more complex physically-based model known as the Snow-Atmosphere-Soil Transfer (SAST) which was developed as part of a land-surface scheme. The SNOW17 uses temperature as the sole index to snowmelt, tracking the accumulation and ablation of the snowpack over time. SAST is a three-layered model that uses standard meteorological variables as input to calculate the energy-balance of the snowpack. The NWS applies the SNOW17 model in a lumped manner, while most physically-based research models are applied on a distributed or point scale. This study will run both models on a watershed scale and route the snowmelt through the NWS rainfall-runoff model (SACSMA) for forecast locations in the Colorado River basin. Streamflow simulations will be evaluated for improved streamflow forecasting and the benefits of adapting features of the SAST model to the NWS snow modeling process assessed.