Snow sublimation is an important component of the snow mass balance, but the spatial and temporal variability of this process is not well understood in mountain environments. This study combines a process‐based snow model (SnowModel) with eddy covariance (EC) measurements to investigate (1) the spatio‐temporal variability of simulated snow sublimation with respect to station observations, (2) the contribution of snow sublimation to the ablation of the snowpack, and (3) the sensitivity and response of snow sublimation to bark beetle‐induced forest mortality and climate warming across the north‐central Colorado Rocky Mountains. EC‐based observations of snow sublimation compared well with simulated snow sublimation at stations dominated by surface and canopy sublimation, but blowing snow sublimation in alpine areas was not well captured by the EC instrumentation. Water balance calculations provided an important validation of simulated sublimation at the watershed scale. Simulated snow sublimation across the study area was equivalent to 28% of winter precipitation on average, and the highest relative snow sublimation fluxes occurred during the lowest snow years. Snow sublimation from forested areas accounted for the majority of sublimation fluxes, highlighting the importance of canopy and sub‐canopy surface sublimation in this region. Simulations incorporating the effects of tree mortality due to bark‐beetle disturbance resulted in a 4% reduction in snow sublimation from forested areas. Snow sublimation rates corresponding to climate warming simulations remained unchanged or slightly increased, but total sublimation losses decreased by up to 6% because of a reduction in snow covered area and duration.