In many parts of the western United States, winter snow acts as a natural reservoir, gradually melting to supply water throughout the year. Predicting how much snow will accumulate and when it will melt is critical for managing water resources — but it’s especially difficult in mountainous areas where weather conditions vary dramatically with elevation. Most models simplify this by assuming a fixed temperature change with elevation, but this can introduce errors. In this study, we tested a more flexible approach that accounts for how temperature and other weather conditions vary hour-by-hour across terrain. We used this method to improve snow simulations in the Sierra Nevada and Cascade mountain ranges. The results showed that the new method improved predictions of air temperature, snow depth, and snow water content — especially during a major snow drought year. As snow droughts become more common with climate change, better snowpack modeling will be essential for water planning. Our approach offers a simple, low-cost way to make existing models more accurate in mountainous regions.