Significant differences in snow deposition, development of the seasonal snow cover, and the timing of melt, occur between two experimental sites located in a headwater sub-drainage of the Reynolds Creek Experimental Watershed (RCEW) in the Owyhee Mountains of southwestern Idaho, USA. Snow, climate and stream discharge data were used to drive and verify a point snow-cover energy and mass balance model, SNOBAL, to evaluate these differences for three water years: 1984, the largest discharge year on record (204% of average), 1992, the smallest discharge year on record (36% of average), and 1999 (140% of average). The simulations showed that wind differences between sites are the most important cause of the snow cover differences between the sites. Though differences do occur in the snow-cover energy balance, these are primarily caused by the fact that reduced precipitation causes melt-out to occur earlier at site 176. This analysis shows the importance of understanding and accounting for variable patterns of snow deposition and redistribution of snow in semiarid regions.