The amount of snow on the ground is mainly controlled by local temperature and precipitation conditions, a fact which is the backbone of numerical snow models in a variety of contexts. Both simple and complex numerical snow models are widely used to produce hydrological predictions, avalanche warnings, and boundary conditions for the atmosphere models that produce weather forecasts. In my work, I investigate the sources of discrepancy among different snow datasets—particularly in how they represent observed historical conditions—and assess the influence of modeling uncertainty and biases in forcing variables.