Snow anomalies in the western United States (U.S.) have been widely investigated by many researchers due to its impact on water availability. This study focuses on how anomalous atmospheric circulation affects the snowpack accumulation in the western U.S. through a comparison of observations and the output of the Community Climate Model version 3 (CCM3, National Center for Atmospheric Research). Our results show that the snow anomalies in the western U.S. correlate with anomalous atmospheric patterns which change the normal precipitation and temperature advection distributions. Further analysis indicates that the midlatitude atmospheric circulation anomalies induced by the El Nino-Southern Oscillation (ENSO) drive the snowfall shifts, and lead to an anomalous distribution of the snowpack in the western U.S. During the warm phase of the ENSO, the snow depth in the Northwest is reduced, while the snow depth in the Southwest is deeper. The cold phase of the ENSO correlates with the higher snow depth in the Northwest, but lower snow depth in the Southwest. Independent from ENSO, the non-linear dynamics-related Pacific-North American (PNA) pattern in the atmosphere is also found to significantly affect snow anomalies in the western U.S. The positive phase of the PNA pattern produces cold temperature advection during snow seasons, which reduces the snowmelt and results in the above normal snowpack in most part of the western U.S., while the negative phase of this pattern generates warm temperature advection, and leads to below normal snowpack. Finally, the performance of CCM3 is evaluated based on the observations, and biases in the model are identified. The information provided in this study strengthens our understanding of climate and water supply variability in the western U.S.